December 2023 - nexa - server-nexa.polito.it

"Posta elettronica": persi 6 Atenei in ~3 anni...
by Damiano Verzulli Jan. 10, 2024

Jan. 10, 2024

Come alcuni ricorderanno, poco meno di tre anni fa --esattamente il 14/02/2021-- fotografai lo stato delle infrastrutture di Posta Elettronica in uso nei vari Atenei del nostro Paese. Il quadro che ne emerse non era particolarmente confortante: * Google: 31 * Microsoft: 16 * Interna all’Ateneo: 19 Oggi (27/12/2023), approfittando di un po' di tempo libero, ho "aggiornato" la fotografica, prendendo in esame *ESATTAMENTE* gli stessi 66 domini di posta analizzati a suo tempo. La situazione aggiornata è la seguente: * Google: 33 * Microsoft: 20 * Interna all’Ateneo: 13 In poco meno di tre anni, abbiamo "perso" 6 atenei/istituzioni: * sissa.it * unica.it * unipa.it * unirc.it * unisannio.it * uniss.it L'articolo integrale --aggiornato con questi ultimi valori-- è disponibile on-line: https://dvblog.soabit.com/la-posta-elettronica-negli-atenei-italiani/ Bye, DV -- Damiano Verzulli e-mail:damiano@verzulli.it --- possible?ok:while(!possible){open_mindedness++} --- "...I realized that free software would not generate the kind of income that was needed. Maybe in USA or Europe, you may be able to get a well paying job as a free software developer, but not here [in Africa]..." -- Guido Sohne - 1973-2008 http://ole.kenic.or.ke/pipermail/skunkworks/2008-April/005989.html

14 25

The EU should support Ireland’s bold move to regulate Big Tech
by Daniela Tafani Jan. 2, 2024

Jan. 2, 2024

The EU should support Ireland’s bold move to regulate Big Tech by Zephyr Teachout and Roger McNamee Dublin, Ireland, was stunned a month ago by riots that transformed its downtown into chaos, the worst rioting in decades, stemming from far-right online rumors about an attack on children.  The riots, like Jan. 6, appear to be a direct outgrowth of the amplification ecosystem supported by social media networks such as TikTok, Google’s YouTube and Meta’s Instagram, which likely keep their European headquarters in Dublin for tax reasons.  Ireland, long ridiculed for bowing to Big Tech, has now come out with a powerful proposal to address the problems of algorithmic amplification. Ireland set up Coimisiún na Meán, a new enforcer, this year to set rules for digital platforms. It has proposed a simple, easily enforceable rule that could change the game: All recommender systems based on intimately profiling people should be turned off by default.  In practice, that means that the big platforms cannot automatically run algorithms that use information about a person’s political views, sex life, health or ethnicity. A person will be able to switch an algorithm on, but those toxic algorithms will no longer be on by default. Users will still have access to algorithmic amplification, but they will have to opt in to get it. Today, algorithms feed each user different information. They derive their power from the trove of personal data that platforms acquire about users, data that enables the identification and exploitation of emotional weak points, all to maximize engagement. Platforms do not acknowledge responsibility for downstream consequences. Some users respond best to cat videos, others to hate speech, disinformation and conspiracy theories. For many, the response to harmful content is involuntary, driven by flight or fight. Either way, users spend more time on the platform, which allows the company to make more money by showing them ads. Artificially amplifying outrage may be lucrative, but it carries a terrible cost. Recommender systems enable to migrate from the fringe to the mainstream. An investigation revealed that Meta’s algorithms were key contributors to the murderous hate that cost thousands of people their lives in Myanmar’s 2017 Rohingya genocide.  Frances Haugen, a whistleblower, revealed in 2021 that Meta had known the danger of its algorithm for years. As long ago as 2016, Meta’s internal research had reported that “64 percent of all extremist group joins are due to our recommendation tools.” It continued: “Our recommendation systems grow the problem.”  When 37,000 people allowed researchers to monitor their YouTube experience, nearly all the nasty videos they had encountered were pushed into their feed by YouTube’s algorithm. An experiment using simulated users found that Facebook, Instagram and X, the platform formerly known as Twitter, recommended antisemitic and conspiracy content to test users as young as 14 years old.   Earlier this year, the surgeon general spoke about the danger of algorithms that promote suicide and self-harm. A recent experiment by Amnesty International proves the point: TikTok’s algorithm recommended videos encouraging suicide to a test user posing as a 13-year-old only an hour after the account was created.  Recommender systems didn’t really take hold until 2010, and we’ve tried trusting the Big Tech platforms, but 13 years into the experiment, with plenty of data showing harm, we cannot trust technology companies to regulate themselves in the public interest.  We do not want the government to be in the business of sorting through what is and is not harmful if amplified. The brilliance of the Ireland model is that it offers a way forward: rules that are content-neutral, giving users control of one critical aspect of their online experience. After years of being a tax haven for Big Tech, Ireland is now offering the world a groundbreaking rule to protect democracy, public health and public safety. In under nine months, Coimisiún na Meán has gone from initial launch to tackling the machine at the heart of the disinformation crisis. The rule is necessary because current European Union regulations aren’t working and the new Digital Services Act is not designed to tackle the core problem. Under the EU’s General Data Protection Regulation, tech firms are already supposed to get a person’s “explicit” (two-step) consent to process inferences about their political views, sexuality, religion, ethnicity or health. Several complaints that the big firms have failed to seek or receive this consent have not been resolved years after being brought. But for over five years, Big Tech’s primary General Data Protection Regulation authority, which is in Ireland, failed to notice or act.  Europe often trumpets its regulatory leadership in the world. But the so-called “Brussels Effect” of other countries heeding its rules began to dissipate when Europe failed to enforce its most famous law: the General Data Protection Regulation. The European Commission is understandably focused on the Digital Services Act, which goes into effect next month, but EU policymakers should welcome the new Irish rules. Coimisiún na Meán’s bold move would ultimately make the Digital Services Act far more successful. Europe and the Irish government are stepping up at last to regulate harmful technology products. Social media may become social again.  https://thehill.com/opinion/technology/4380369-the-eu-should-support-irelan…

3 3

Re: [nexa] AI/LLM a trazione europea...
by Giuseppe Attardi Dec. 30, 2023

Dec. 30, 2023

Mistral ha ricevuto un primo investimento di 120 milioni per sviluppare sistemi di AI, il governo italiano ha investito 120 milioni per l’acquisto di Leonardo (raddoppiando i 120 della Commissione Europea). Qual è stata la decisione più furba? BTW, Leonardo è semplicemente un prodotto commerciale di Athos, industria francese. A quel tempo, consultando il MISE, cercai di suggerire che il CINECA facesse uno sforzo per progettare un’architettura di computer dedicato all’AI. Per gennaio sono stato convocato di nuovo per un’audizione sull’AI alla Camera e sono andato a rileggere il mio intervento analogo al Senato del 2020. Dicevo cose che state dicendo adesso: che ci vorrebbe un’iniziativa di dimensione europea sull’AI, la chiamavo un CERN for AI e citavo le capacità di GPT-3 come indicatore di possibili traguardi. https://www.senato.it/Leg18/3557?indagine=1809 Comunque, Mixtral7B è un primo risultato promettente. Il confronto con gli altri modelli SotA però è un po’ deludente. https://arxiv.org/pdf/2312.11444.pdf Per accrescere la dimensione del modello, occorre arrivare a utilizzare oltre un miliardo di MIPS, che nemmeno Allen Artificial Intelligenge riesce a permettersi per allenare il loro modello OLMo. Ossia, Leonardo andrebbe interamente dedicato ad allenare un LLM per circa un mese. Qualcuno deve pagare per tutte quelle risorse e per l’energia elettrica. Inoltre la strada da percorrere porta verso i modelli multimodali, tipo Gemini, e lì il compito si fa ancora più arduo. Se i francesi non fossero così sciovinisti e lo stesso i tedeschi e gli italiani, si dovrebbero mettere tutti insieme per lavorare ad un obiettivo comune. OpenAI ha 770 dipendenti, circa 200 hanno firmato l’articolo su GPT-4. La foto del team di Mistral mostra una dozzina di persone. BTW, la Commissione ha annunciato un finanziamento di 100 milioni per un progetto sui LLM nell’ambito dell’iniziativa Interesting Projects of Common European Interest, di cui è capofila Tiscali (https://ec.europa.eu/commission/presscorner/detail/en/IP_23_6246) C’è poi un finanziamento di 4 milioni (H4LM) per fare una raccolta di testi multilingue per allenare LLM (https://hplt-project.org/) — Beppe > On 27 Dec 2023, at 09:06, nexa-request(a)server-nexa.polito.it wrote: > From: Damiano Verzulli <damiano(a)verzulli.it> > To: "nexa(a)server-nexa.polito.it" <nexa(a)server-nexa.polito.it> > Subject: [nexa] AI/LLM a trazione europea... > Message-ID: <42dcc85e-6452-4255-9ddd-7f67a1e6fdae(a)verzulli.it> > Content-Type: text/plain; charset="utf-8"; Format="flowed" > > In questo gruppo si parla spesso di AI/LLM e... spesso ho notato mani > alzate a mo' di: "Eh! Ma servono 'risorse'... servono 'soldi'.... > servono 'infrastrutture'... perché senza, non si va da nessuna parte". > > Ricordo, ad esempio, un paio di thread incentrati sul fatto che UniPisa > avesse investito ~1M€ (se ricordo bene...) per acquistare un paio di > "cluster" [di cui almeno uno acquistato con fondi PNRR], specializzati > sul GPU-computing.... > > Bene... dico questo perché qualche settimana fa ho incrociato un POST di > Vincenzo Di Nicola [1] il quale evidenziava come Mistral [2], una > startup francese partita a maggio 2023, fosse valutata 2G€ e... stesse > rapidissimamente diventando una entita' di spicco nel panorama AI mondiale. > > Ripeto: è una startup francese. > > La cosa che piu' di tutte mi ha colpito, pero', è un'altra. Vincenzo > scriveva: "...E Mistral, in pochi mesi, ha gia’ prodotto un LLM > addestrato velocemente grazie al nostro supercomputer Leonardo del > Cineca..." > > Ecco: finalmente trovavo una evidenza concreta di un utilizzo di risorse > --finanziate e gestite dal nostro Paese-- che andasse a beneficio di > altre risorse/entita'... mainstream e, soprattutto, made-in-Europe. > > La notizia mi era nuova... e quindi ho chiesto a Vincenzo una conferma. > Oltre a confermarla, mi ha girato un ulteriore riferimento (paywalled > [3]) e, successivamente è anche andato a farsi un giro in Cineca [4] > > Cercando in giro, ho trovato quest'altra fonte ([5] in francese) che > cita sia Leonardo che EuroHPC... > > Insomma: a dispetto di quanto si pensi nei corridoi di molti nostri > Atenei... è perfettamente possibile utilizzare Leonardo per addrestare > un LLM... e tra l'altro in tempi rapidissimi. > > Mi chiedo perché dei Francesi abbiano potuto farlo... e degli Italiani > no. Ma è un dettaglio :-) > > Un saluto, > DV > > [1] > https://www.linkedin.com/posts/dinicola_openai-rival-mistral-nears-2-billio… > [2] https://mistral.ai/ > [3] > https://www.theinformation.com/articles/why-do-we-need-an-openai-of-europe > [4] > https://www.linkedin.com/posts/dinicola_supercomputer-francescoubertini-cin… > [5] https://www.silicon.fr/mistral-ai-llm-471793.html > > -- > Damiano Verzulli > e-mail:damiano@verzulli.it > --- > From: Damiano Verzulli <damiano(a)verzulli.it> > To: "nexa(a)server-nexa.polito.it" <nexa(a)server-nexa.polito.it> > Subject: [nexa] AI/LLM a trazione europea... > Message-ID: <42dcc85e-6452-4255-9ddd-7f67a1e6fdae(a)verzulli.it> > Content-Type: text/plain; charset="utf-8"; Format="flowed" > > In questo gruppo si parla spesso di AI/LLM e... spesso ho notato mani > alzate a mo' di: "Eh! Ma servono 'risorse'... servono 'soldi'.... > servono 'infrastrutture'... perché senza, non si va da nessuna parte". > > Ricordo, ad esempio, un paio di thread incentrati sul fatto che UniPisa > avesse investito ~1M€ (se ricordo bene...) per acquistare un paio di > "cluster" [di cui almeno uno acquistato con fondi PNRR], specializzati > sul GPU-computing.... > > Bene... dico questo perché qualche settimana fa ho incrociato un POST di > Vincenzo Di Nicola [1] il quale evidenziava come Mistral [2], una > startup francese partita a maggio 2023, fosse valutata 2G€ e... stesse > rapidissimamente diventando una entita' di spicco nel panorama AI mondiale. > > Ripeto: è una startup francese. > > La cosa che piu' di tutte mi ha colpito, pero', è un'altra. Vincenzo > scriveva: "...E Mistral, in pochi mesi, ha gia’ prodotto un LLM > addestrato velocemente grazie al nostro supercomputer Leonardo del > Cineca..." > > Ecco: finalmente trovavo una evidenza concreta di un utilizzo di risorse > --finanziate e gestite dal nostro Paese-- che andasse a beneficio di > altre risorse/entita'... mainstream e, soprattutto, made-in-Europe. > > La notizia mi era nuova... e quindi ho chiesto a Vincenzo una conferma. > Oltre a confermarla, mi ha girato un ulteriore riferimento (paywalled > [3]) e, successivamente è anche andato a farsi un giro in Cineca [4] > > Cercando in giro, ho trovato quest'altra fonte ([5] in francese) che > cita sia Leonardo che EuroHPC... > > Insomma: a dispetto di quanto si pensi nei corridoi di molti nostri > Atenei... è perfettamente possibile utilizzare Leonardo per addrestare > un LLM... e tra l'altro in tempi rapidissimi. > > Mi chiedo perché dei Francesi abbiano potuto farlo... e degli Italiani > no. Ma è un dettaglio :-) > > Un saluto, > DV > > [1] > https://www.linkedin.com/posts/dinicola_openai-rival-mistral-nears-2-billio… > [2] https://mistral.ai/ > [3] > https://www.theinformation.com/articles/why-do-we-need-an-openai-of-europe > [4] > https://www.linkedin.com/posts/dinicola_supercomputer-francescoubertini-cin… > [5] https://www.silicon.fr/mistral-ai-llm-471793.html > > -- > Damiano Verzulli > e-mail:damiano@verzulli.it > ---

2 1

Re: [nexa] AI/LLM a trazione europea...
by Giuseppe Attardi Dec. 29, 2023

Dec. 29, 2023

1 0

Il New York Times fa causa a OpenAI e Microsoft per violazione del copyright
by Antonio Dec. 28, 2023

Dec. 28, 2023

https://nytco-assets.nytimes.com/2023/12/NYT_Complaint_Dec2023.pdf Buona lettura (sono 69 pagine) Qui https://fingfx.thomsonreuters.com/gfx/legaldocs/lgvdlrybjvo/OPENAI%20COPYRI… un'altra denuncia da parte di autori che, "complessivamente, hanno vinto tre premi Pulitzer e dedicato decenni della loro vita e somme ingenti per la ricerca, l'affinamento, la scrittura e il perfezionamento delle loro opere". Ciao, Antonio

1 0

AI/LLM a trazione europea...
by Damiano Verzulli Dec. 27, 2023

Dec. 27, 2023

In questo gruppo si parla spesso di AI/LLM e... spesso ho notato mani alzate a mo' di: "Eh! Ma servono 'risorse'... servono 'soldi'.... servono 'infrastrutture'... perché senza, non si va da nessuna parte". Ricordo, ad esempio, un paio di thread incentrati sul fatto che UniPisa avesse investito ~1M€ (se ricordo bene...) per acquistare un paio di "cluster" [di cui almeno uno acquistato con fondi PNRR], specializzati sul GPU-computing.... Bene... dico questo perché qualche settimana fa ho incrociato un POST di Vincenzo Di Nicola [1] il quale evidenziava come Mistral [2], una startup francese partita a maggio 2023, fosse valutata 2G€ e... stesse rapidissimamente diventando una entita' di spicco nel panorama AI mondiale. Ripeto: è una startup francese. La cosa che piu' di tutte mi ha colpito, pero', è un'altra. Vincenzo scriveva: "...E Mistral, in pochi mesi, ha gia’ prodotto un LLM addestrato velocemente grazie al nostro supercomputer Leonardo del Cineca..." Ecco: finalmente trovavo una evidenza concreta di un utilizzo di risorse --finanziate e gestite dal nostro Paese-- che andasse a beneficio di altre risorse/entita'... mainstream e, soprattutto, made-in-Europe. La notizia mi era nuova... e quindi ho chiesto a Vincenzo una conferma. Oltre a confermarla, mi ha girato un ulteriore riferimento (paywalled [3]) e, successivamente è anche andato a farsi un giro in Cineca [4] Cercando in giro, ho trovato quest'altra fonte ([5] in francese) che cita sia Leonardo che EuroHPC... Insomma: a dispetto di quanto si pensi nei corridoi di molti nostri Atenei... è perfettamente possibile utilizzare Leonardo per addrestare un LLM... e tra l'altro in tempi rapidissimi. Mi chiedo perché dei Francesi abbiano potuto farlo... e degli Italiani no. Ma è un dettaglio :-) Un saluto, DV [1] https://www.linkedin.com/posts/dinicola_openai-rival-mistral-nears-2-billio… [2] https://mistral.ai/ [3] https://www.theinformation.com/articles/why-do-we-need-an-openai-of-europe [4] https://www.linkedin.com/posts/dinicola_supercomputer-francescoubertini-cin… [5] https://www.silicon.fr/mistral-ai-llm-471793.html -- Damiano Verzulli e-mail:damiano@verzulli.it --- possible?ok:while(!possible){open_mindedness++} --- "...I realized that free software would not generate the kind of income that was needed. Maybe in USA or Europe, you may be able to get a well paying job as a free software developer, but not here [in Africa]..." -- Guido Sohne - 1973-2008 http://ole.kenic.or.ke/pipermail/skunkworks/2008-April/005989.html

2 1

AI cannot patent inventions, UK Supreme Court confirms
by don Luca Peyron Dec. 27, 2023

Dec. 27, 2023

A conferma di quanto già dicevamo sin dall'inizio Buona lettura e buon tempo di Natale dl https://www.bbc.com/news/technology-67772177?utm_source=www.mindstream.news… _________________________ don Luca Peyron Pastorale Universitaria - Apostolato Digitale Arcidiocesi di Torino www.universitari.to.it via XX settembre 83, Torino tel. 011 5156239

1 0

168° Mercoledì di Nexa | 10 gennaio 2024, ore 17.00
by Valeria Bergantino Dec. 21, 2023

Dec. 21, 2023

Gentilissime, gentilissimi, Vi invitiamo a partecipare al *168° Mercoledì di Nexa*, che si terrà *mercoledì 10 gennaio*, alle ore 17.00, con un incontro dal titolo /*"L’architetto e l’oracolo. Forme digitali del sapere da Wikipedia a ChatGPT"*/. Ospite dell'incontro: *Gino Roncaglia* (Università degli Studi Roma Tre)/*.*/ _L'incontro si terrà IN PRESENZA e ONLINE._ _SEDE FISICA_ dell'incontro: Centro Nexa su Internet e Società, Politecnico di Torino, Via Boggio 65/a, Torino (1° piano). Per accedere alla sala si raccomanda di suonare al citofono *Portineria* e di seguire le indicazioni segnalate lungo il percorso. QUI <https://nexa.polito.it/contatti> maggiori informazioni su come raggiungerci. _STANZA VIRTUALE_ dell'incontro: https://didattica.polito.it/VClass/NexaEvent Di seguito maggiori dettagli: Se non visualizzi correttamente questo messaggio clicca qui <https://nexa.polito.it/mercoledi-168> NEXA 168° Mercoledì di Nexa Mercoledì 10 gennaio 2024, ore 17.00 - 19.00 Politecnico di Torino https://nexa.polito.it/mercoledi-168 /L'architetto e l'oracolo./ /Forme digitali del sapere da Wikipedia a ChatGPT/ *GINO RONCAGLIA (Università degli Studi Roma Tre)* *L'INCONTRO SI TERRÀ IN PRESENZA E ONLINE* *SEDE FISICA*: Centro Nexa su Internet e Società, Politecnico di Torino, Via Boggio 65/a, Torino (1° piano). Suonare al citofono *Portineria* - Seguire le indicazioni segnalate dai cartelli lungo il percorso. (Per maggiori informazioni su come raggiungerci clicca QUI <https://nexa.polito.it/contatti>) *STANZA VIRTUALE*: https://didattica.polito.it/VClass/NexaEvent Al tradizionale modello architettonico e sistematico di organizzazione delle conoscenze, al quale fanno riferimento strumenti come le enciclopedie, i sistemi di catalogazione e metadatazione, o l’organizzazione disciplinare tipica di molti processi formativi, si contrappone oggi un modello assai meno strutturato e organizzato, che cresce e si sviluppa in forme non sempre prevedibili: quello delle *nuove intelligenze artificiali generative, che operano in base a un paradigma statistico-probabilistico e dunque non strettamente deterministico*. Le intelligenze artificiali generative sono state paragonate a ‘pappagalli stocastici’, ma sono forse meglio descrivibili come oracoli predittivi: l’intervento discuterà il rapporto fra i due paradigmi, ne proporrà una rapida ricostruzione storica (con particolare riferimento al passaggio dall’iniziale contrapposizione fra IA forte e IA debole nella ‘Good Old-Fashioned Artificial Intelligence’ della seconda metà del secolo scorso, a quella attuale fra i concetti di Narrow AI e General AI) e fornirà una sintetica introduzione al funzionamento delle *reti neurali generative*. Nell’ultima parte saranno poi proposte alcune considerazioni sulla situazione attuale e sulle prevedibili evoluzioni future dell’intelligenza artificiale generativa, in particolare in direzione della *multi-modalità* e dell’*embodiment*. BIOGRAFIA e informazioni supplementari: [mercoledì168] *GINO RONCAGLIA* è professore associato presso l’Università Roma Tre, dove insegna Informatica umanistica, Filosofia dell’informazione ed Editoria digitale. Autore di numerosi saggi sul settore dei nuovi media (con particolare riferimento al futuro del libro e della lettura), è fra gli esperti che hanno partecipato alla stesura del Piano Nazionale Scuola Digitale e collabora da tempo con RAI Cultura. Proprio per RAI Cultura ha realizzato durante l’emergenza COVID-19 i programmi /Scuola @ Casa/ (10 puntate di introduzione alla didattica a distanza) e /Scuola @ Casa News/ (40 puntate di informazioni e notizie destinate al mondo della scuola), e ha curato il programma settimanale /Scuola News/. È membro del Consiglio scientifico della Fondazione scuola beni e attività culturali del Ministero della Cultura, e professore onorario della Universidad Nacional de Villa María in Argentina. Fra i suoi libri più recenti, /L’architetto e l’oracolo. Forme digitali del sapere da Wikipedia a ChatGPT/(Laterza 2023), /L’età della frammentazione. Cultura del libro e scuola digitale/ (Laterza 2018, seconda edizione accresciuta, Laterza 2020) e /La quarta rivoluzione. Sei lezioni sul futuro del libro/ (Laterza 2010). *Letture consigliate:* * Roncaglia, G. (2023), /L’architetto e l’oracolo. Forme digitali del sapere da Wikipedia a ChatGPT/, Laterza * Tegmark, M. (2018), /Vita 3.0. Essere umani nell'era dell'intelligenza artificiale/, Raffaello Cortina Editore * Roncaglia, G. (4 maggio 2023), /Intelligenze artificiali generative e mediazione informativa: una introduzione/, Biblioteche Oggi Trends, N.1, disponibile al LINK <https://www.bibliotecheoggitrends.it/it/articolo/2839/intelligenze-artifici…> Che cosa sono il Centro Nexa e i cicli di incontri “Mercoledì di Nexa” e “Nexa Lunch Seminar” <https://www.facebook.com/nexa.center/> <https://twitter.com/nexacenter> <https://www.youtube.com/user/NexaCenter> <https://www.instagram.com/nexa_center/> <https://www.linkedin.com/company/3054864/admin/> #nexawednesday <https://twitter.com/search?q=%23nexawednesday&src=typd> #GinoRoncaglia <https://twitter.com/roncaglia> #UniversitaRomaTre <https://twitter.com/UnivRoma3> Il Centro Nexa su Internet & Società del Politecnico di Torino (Dipartimento di Automatica e Informatica), fondato nel 2006, è un centro di ricerca interdisciplinare che, in collaborazione con l’Università di Torino (in particolare il Dipartimento di Giurisprudenza), studia le tecnologie digitali e il loro rapporto con la società. Maggiori informazioni all'indirizzo: http://nexa.polito.it/about. Durante i “Mercoledì di Nexa”, che si tengono *ogni 2° mercoledì del mese alle ore 17 in punto*, il Centro Nexa su Internet e Società apre le sue porte non solo agli esperti e a tutti coloro i quali lavorano con Internet, ma anche a semplici appassionati e cittadini. Il ciclo di incontri intende approfondire, con un linguaggio preciso ma accessibile, i temi legati alla Rete: “intelligenza artificiale”, reti sociali, software libero, capitalismo della sorveglianza, neutralità della rete, libertà di espressione, privacy, condivisione di file, "big data" e "open data", “smart cities”, e molto altro. Al centro della maggior parte degli incontri un ospite pronto a dialogare con i direttori del Centro Nexa, il Prof. Juan Carlos De Martin del Politecnico di Torino, i Proff. Marco Ricolfi e Maurizio Borghi dell'Università di Torino, lo staff, i Fellows del Centro Nexa e tutti i presenti. Maggiori informazioni sui Mercoledì di Nexa, incluso un elenco di tutti i “Mercoledì” passati, sono disponibili all'indirizzo: http://nexa.polito.it/mercoledi. Le registrazioni degli incontri sono disponibili qui: https://nexa.polito.it/audio-video. Si segnala inoltre che dal maggio 2012 *ogni 4° mercoledì* del mese *dalle ore 13 alle ore 14* il Centro Nexa organizza anche i "*Nexa Lunch Seminar*". Una lista di tutti i “Lunch Seminar” passati è disponibile all'indirizzo: http://nexa.polito.it/lunch-seminars. See our events calendar <http://nexa.polito.it/events> if you're curious about future luncheons, discussions, lectures, and conferences not listed in this email. Our events are free and open to the public, unless otherwise noted. Responsabile Comunicazione Centro Nexa su Internet & Società: *Valeria Bergantino*, tel: +39 011 090 7219, Mob: +39 347 344 3585, valeria.bergantino(a)polito.it. Maggiori informazioni sui Mercoledì di Nexa e i Nexa Lunch Seminar, sono disponibili all'indirizzo: http://nexa.polito.it/events. Weekly Events Newsletter. Sign up <http://nexa.polito.it/mailing-lists> to receive this newsletter if this email was forwarded to you. To manage your subscription preferences, please click here <https://server-nexa.polito.it/cgi-bin/mailman/listinfo/nexa>. Connect & get involved: Jobs, internships, and more <http://nexa.polito.it/get-involved>. Nexa Center for Internet and Society Newsletter Cordiali saluti, -- Valeria Bergantino Communication Officer Nexa Center for Internet & Society Politecnico di Torino - DAUIN Via Pier Carlo Boggio, 65/A - 10138 Torino web: https://nexa.polito.it/ mail: valeria.bergantino(a)polito.it <mailto:valeria.bergantino@polito.it> tel: 3473443585

1 0

Presunzioni delle macchine e consenso dell'interessato
by Daniele Imbruglia Dec. 19, 2023

Dec. 19, 2023

Salve nexian*, mi permetto di condividere una recente pubblicazione, allegata, anche per ringraziare i tanti input ricevuti da questa magnifica mailing list. Grato per ogni commento, auguro a tutti buone vacanze, Daniele -- dr. Daniele Imbruglia Ricercatore in diritto privato Dipartimento di Diritto ed economia delle attività produttive Facoltà di Economia - Università di Roma "La Sapienza" Via del Castro Laurenziano, 9 - 00161 Roma RM https://uniroma1.academia.edu/DanieleImbruglia

1 0

E-books are fast becoming tools of corporate surveillance
by Alberto Cammozzo Dec. 18, 2023

Dec. 18, 2023

As the Internet Archive appeals a court decision blocking alternatives to surveillance-ridden digital book licenses, a new report reveals that the world’s largest publisher may be selling readers’ intimate personal data to the highest bidder <https://www.fastcompany.com/90996547/e-books-are-fast-becoming-tools-of-cor…> Three in ten <https://www.pewresearch.org/short-reads/2022/01/06/three-in-ten-americans-n…> Americans read digital books. Whether they’re accessing online textbooks or checking out the latest bestselling e-book from the public library, the majority of these readers are subject to both the greed of Big Publishing and the priorities of Big Tech. In fact, Amazon’s Kindle held 72% <https://wordsrated.com/amazon-kindle-e-book-and-kindle-unlimited-statistics/> of the e-reader market in 2022. And if there’s one thing we know about Big Tech companies like Amazon <https://www.businessinsider.com/amazon-empire-of-surveillance-leveraging-mo…>, their real product isn’t the book. It’s the user data. Major publishers are giving Big Tech free rein to watch what you read and where, including books on sensitive topics, like if you check out a book on self care after an abortion. Worse, tech and publishing corporations are gobbling up data beyond your reading habits—today, there are no federal laws to stop them from surveilling people who read digital books across the entire internet. Reader surveillance is a deeply intersectional threat, according to a congressional letter <https://www.fightforthefuture.org/news/2023-12-07-25-human-rights-organizat…> issued last week from a coalition of groups whose interests span civil rights, anti-surveillance, anti-book ban, racial justice, reproductive justice, LGBTQ+, immigrant, and antimonopoly. Our letter calls on federal lawmakers to investigate the harms of tech and publishing corporations’ powerful hold over digital book access. This investigation is an essential first step to revive the right to read without fear of having your interests used against you. Because unfortunately, that right is on life support when it comes to digital books. Right now, there’s a concerted effort from anti-abortion lawmakers and vigilantes to discover who is getting or supporting an abortion, and data brokers have already been sued <https://www.ftc.gov/news-events/news/press-releases/2022/08/ftc-sues-kochav…> for packaging up data on visitors to reproductive health centers. But the U.S. still has no meaningful federal data privacy <https://www.eff.org/deeplinks/2023/04/digital-privacy-legislation-civil-rig…> protections to turn off the tap of intimate personal data. Whether you’re traveling for abortion care, legally accessing it at home <https://www.plancpills.org/>, experiencing a miscarriage <https://www.npr.org/2022/07/03/1109015302/abortion-prosecuting-pregnancy-lo…>, or finding support resources <https://mahotline.org/> for a loved one, worrying that reading a book might betray private medical information should be absurd. Yet in today’s post-Roe landscape of legal, social, and health threats around pregnancy, we need to be concerned that Big Publishing’s new hunger for reader data might put abortion seekers at risk of <https://www.eff.org/deeplinks/2023/04/digital-privacy-legislation-civil-rig…> criminalization and violence <https://www.insider.com/texas-woman-killed-boyfriend-getting-abortion-2023-5>. Such /1984/-esque threats woven into the lives of everyday people are particularly insidious. The average abortion seeker does not know <https://jezebel.com/nebraska-mom-pleads-guilty-to-giving-abortion-pills-to-…> how to navigate safely in a surveillance state—and knowing that they may be surveilled causes people to not access <https://societyfp.org/wp-content/uploads/2023/06/WeCount-Release-June-2023.…> the reproductive or abortion care they need, to not seek the mental and emotional support they require, and to delay care until further into an unwanted pregnancy. All of these choices endanger a pregnant person. In the age of artificial intelligence, the ability to analyze unfathomably detailed data on individual people, create reports and inferences about those people, and use the whole lot of it to train AI models is constantly improving. The incentives to exploit the data of readers are the strongest they have ever been. Big Publishing is clearly seeing nothing but dollar signs <https://www.newyorker.com/news/annals-of-communications/an-app-called-libby…> as apps like Hoopla <https://apps.apple.com/us/app/hoopla-digital/id580643740> gobble up identity-linked data on readers—and so it would be natural to put our hope in public libraries, which view patron privacy as a fundamental right essential to a functioning democracy. In the human rights community, libraries’ resistance against government surveillance <https://slate.com/technology/2015/06/usa-freedom-act-before-snowden-librari…> under the Patriot Act is legendary. Unfortunately, Big Publishing has sued to stop libraries from loaning surveillance-free digital books—winning a lower court judgment that the nonprofit Internet Archive <https://www.battleforlibraries.com/> is set to appeal before the year is out. Unless that judgment is overturned or new laws are passed, libraries have no alternative but to license digital books that are likely to be riddled with spyware. We know less about surveillance at public libraries because, as a November report <https://sparcopen.org/our-work/big-deal-knowledge-base/confidentiality-clau…> from the Scholarly Publishing and Academic Resources Coalition details, Big Publishing has been increasingly sneaky about privacy and surveillance in their library contracts. This is a play right out of Silicon Valley’s handbook: to hide bad behavior with unaccountable external links or NDAs <https://fortune.com/2019/04/29/silicon-valley-nda/> that prohibit libraries from warning their patrons. Without laws to stop them, it’s reasonable to expect that popular library apps like Hoopla and Libby are hiding similar behavior behind legal smokescreens. Already, the absurdity of Amazon Kindle’s data collection is well documented <https://www.theguardian.com/technology/2020/feb/03/amazon-kindle-data-readi…> and a source of Amazon’s overarching monopsony power <https://jacobin.com/2022/11/cory-doctorow-chokepoint-capitalism-monopoly-te…> in the book market. With libraries facing legal annihilation <https://www.theatlantic.com/ideas/archive/2023/03/publishers-librarians-ebo…> from every direction <https://www.tiktok.com/@fightfortheftr/video/7267303386683297067> if they attempt to carve out surveillance-free spaces for digital books, the future of reading is in the eleventh hour. Lawmakers must immediately launch an investigation to protect not only abortion patients, but all readers throughout this country. ------------------------------------------------------------------------ /Lia Holland is a fiction author and campaigns and communications director at digital rights organization Fight for the Future, where they focus on creator’s rights and emerging technologies./ /Jade Pfaefflin Bounds is a reproductive and racial justice educator, doula, and digital rights organizer with Fight for the Future. His work converges and orients towards collective liberation./ ------------------------------------------------------------------------ /Correction: An earlier version of this op-ed claimed the publisher Elsevier was tracking and selling data to third parties, a claim Elsevier rejects. Such claims have since been removed from the /article/. In a statement, a spokesperson says: “We take the issue of data privacy and security very seriously and maintaining the trust of the people who use our services is critically important to us. Our data privacy practices are transparent, comply with the law and operate within a clear ethical framework. We support the right of individuals to know and manage how their personal information is collected and used, as articulated in our Privacy Principles and Privacy Policy and demonstrated by our user-facing Privacy Center and cookie management platforms.”/

1 0

Articolo "Libertà di ricerca, licenze libere e sistemi d’intelligenza artificiale"
by Marco Ciurcina Dec. 16, 2023

Dec. 16, 2023

Condivido un mio articolo sul tema in oggetto, pubblicato qualche settimana orsono (forse è d'interesse per qualcuno in lista) https://www.rivisteweb.it/doi/10.53227/108473 Abstract After years in which many developments in artificial intelligence systems were the outcome of scientific research and were made available under free licences, the recent success of some generative artificial intelligence systems, which were developed using large computing capacities and quantities of data, leads to the question whether conditions are emerging in which a few players will be able to provide artificial intelligence systems, concentrating economic (but not only) power. But the development of artificial intelligence systems continues. Policies are therefore useful that, by exploiting the regulatory framework (favourable to the development of transparent systems and to the scientific research), make data and computing resources available to researchers in order to foster the free availability, study and understanding of artificial intelligence systems.To achieve this, it is certainly useful to make use of the methods already adopted for free software (which, on the one hand, have led to the generation of digital commons and, on the other, are usefully employed by companies). But this is not enough: artificial intelligence systems do not consist of software alone and require computing and data resources. It is supportive of the European regulatory framework which allows it to adopt policies that enhance the scientific research in the development of artificial intelligence systems thus balancing fundamental rights with conflicting objectives: freedom of enterprise, freedom of scientific research and protection of personal data. L'articolo è in italiano. m.c.

1 0

How Couples Meet and Stay Together (HCMST)
by Luigi Scorca Dec. 16, 2023

Dec. 16, 2023

https://data.stanford.edu/hcmst?fbclid=IwAR2uBvRZ9qf6CSy70BBOUFxoEQuD070nrc… Interessante come l'evoluzione sia stata velocissima..

1 0

EU updates product liability regime to include, software, Artificial Intelligence
by maurizio lana Dec. 15, 2023

Dec. 15, 2023

spesso tra le ragioni istitutive dell'UE quella della creazione del "mercato unico" viene dismissed come una svendita delle idealità a vantaggio della ragione economica. ma qui (product liability regime includes software, Artificial Intelligence) si vede che nell'Antropocene i prodotti non sono giusto delle 'inerti pietre grigie' ma concorrono a dare forma all'ambiente umano e regolare i prodotti significa prendere posizione sui principi. (ok ho scritto Antropocene ma potevo anche dire con intento puramente descrittivo "società del capitalismo"). questo della responsabilità civile dei product manufacturers e degli economic operators di sistemi di IA era un principio già espresso in pressoché tutti i precedenti documenti normativi e legislativi dell'UE che trattavano di IA. a me pare una buona cosa perché esclude agency dei sistemi di IA e di conseguenza impedisce quelle fantasiose cose come responsabilità del sistema intelligente (?). Maurizio Il 15/12/23 12:00, Daniela Tafani <daniela.tafani(a)unipi.it> ha scritto: > Date: Thu, 14 Dec 2023 13:20:57 +0000 > From: Daniela Tafani<daniela.tafani(a)unipi.it> > To:"nexa(a)server-nexa.polito.it" <nexa(a)server-nexa.polito.it> > Subject: [nexa] EU updates product liability regime to include > software, Artificial Intelligence > > EU updates product liability regime to include software, Artificial Intelligence > By Luca Bertuzzi | > > EU policymakers reached a political agreement on Thursday (14 December) on legislation to bring the bloc’s product liability regime in line with technological developments, notably covering digital products like software, which includes Artificial Intelligence. > > The Product Liability Directive (PLD) provides people who have suffered material damage from a defective product with the legal basis to sue the relevant economic operators and seek compensation. > > The product manufacturers will be liable for defectiveness resulting from a component under its control, which might be tangible, intangible, or a related service, like the traffic data of a navigation system. > > Defectiveness > A product is deemed defective when it does not provide the safety a person is entitled to expect based on the reasonable foreseeable use, legal requirements, and the specific needs of the group of users for whom the product is intended. > > One of the elements considered when assessing defectiveness is the capacity of the product to continue to learn and acquire new features or knowledge; wording meant to cover Artificial Intelligence technology based on machine learning techniques. > > Damages > Under the PLD, material damage includes death, personal injury, psychological harm, and destruction of property. At the same time, national liability regimes might still regulate compensation for non-material damages, such as those resulting from discrimination. > > The concept of damage also includes the loss or corruption of data that is not used exclusively for professional purposes. The European Parliament suggested introducing a threshold for the value of the data to qualify as a material loss, but this idea was rejected in the talks with the Council. > > Disclosure of evidence > The PLD introduces the possibility for claimants who have presented sufficient evidence to request that a national court mandates the defendant disclose the evidence at its disposal as long as necessary and proportionate. > > Conversely, the Parliament proposed allowing the defendant to ask for the disclosure of evidence at the disposal of the claimant under the same conditions. > > The disclosing party might request the national court to take the necessary measures to protect the confidentiality of disclosed information including trade secrets. > > Burden of proof > Claimants must prove the product’s defectiveness, the damage suffered, and the causal link between the two. However, under certain conditions, the defectiveness of the product will be assumed, and it will be on the defendant to disprove it. > > The conditions include cases where the defendant fails to disclose relevant evidence, the claimant demonstrates that the product does not comply with mandatory product safety requirements, or an obvious malfunction causes the damage. > > The defectiveness is also presumed in cases where the claimant faces excessive difficulties due to technical or scientific complexity to prove the defectiveness of the product or the causal link, or the claimant proves that it is likely that the product is defective or that there is a causal link. > > Open source software > The Directive will not apply to free and open-source software developed or supplied outside a commercial activity. The liability rules apply when the software is supplied in exchange for a price or personal data used for anything other than improving the software’s security or compatibility. > > Compensation fund > The European Parliament included the possibility of EU countries using existing or new national sectorial compensation schemes for victims of defective products who fail to obtain compensation because the economic operator is insolvent or no longer exists. > > The text specifies that the member states should preferably avoid funding the schemes with public money. > > Liability exemption > The Parliament pushed a liability exemption for manufacturers of software components to a defective product that were micro or small enterprises when they placed that software in the market, provided that another economic operator is liable. > > However, this measure was left for contractual arrangements between the small company and the other party. > > Right of recourse > Parliamentarians included the principle that whenever more than one economic operator is deemed liable for the same damage and is ordered to pay compensation, each one will have the right of recourse against the others. > > Risk development > The Council added the possibility for EU countries to maintain existing measures in their legal systems that make economic operators liable even if they can prove that they could not know the product’s defectiveness based on the state of scientific and technical knowledge. > > Limitation periods > Claimants will have three years to initiate proceedings following the damage. The right for an injured person to seek compensation under this directive expires after ten years, except for latent injuries that are slow to manifest, for which the timeframe is 25 years. > > Timeline > The PLD will apply to all products placed on the EU market 24 months after it enters into force. EU countries will have until then to transpose the directive into national law. > > [Edited by Zoran Radosavljevic] > > L'articolo è in questo luogo occhiuto: > https://www.euractiv.com/section/digital/news/eu-updates-product-liability-… > > ------------------------------ > > Message: 2 > Date: Thu, 14 Dec 2023 14:40:08 +0000 > From: Daniela Tafani<daniela.tafani(a)unipi.it> > To:"nexa(a)server-nexa.polito.it" <nexa(a)server-nexa.polito.it> > Subject: [nexa] A proposito di prodotti pericolosi e non funzionanti: > Tesla recalls nearly all 2 million of its vehicles on US roads > Message-ID:<4b954e94abc14570919bcb2e9f94bae4(a)unipi.it> > Content-Type: text/plain; charset="Windows-1252" > > Tesla recalls nearly all 2 million of its vehicles on US roads > Chris Isidore, CNN > > Tesla is recalling nearly all 2 million of its cars on US roads to limit the use of its Autopilot feature following a two-year probe by US safety regulators of roughly 1,000 crashes in which the feature was engaged. > > The limitations on Autopilot serve as a blow to Tesla’s efforts to market its vehicles to buyers willing to pay extra to have their cars do the driving for them. > > The National Highway Traffic Safety Administration said the Autopilot system can give drivers a false sense of security and be easily misused in certain dangerous situations when a Tesla’s technology may be unable to safely navigate the road. The over-the-air software update will give Tesla drivers more warnings when they are not paying attention to the road while the Autopilot’s “Autosteer” function is turned on. Those notifications will remind drivers to keep their hands on the wheel and pay attention to the road, according to a statement from NHTSA > > After the recall, Teslas with Autosteer turned on will more routinely check on the driver’s attention level – and may disengage the feature – when the software determines the driver isn’t paying attention, when the car is approaching traffic controls, or when it’s off the highway when Autosteer alone isn’t sufficient to drive the car. > > The recall was disclosed in a letter to Tesla posted by NHTSA, which said that Tesla had agreed to the software update starting on Tuesday that will limit the use of the Autosteer feature if a driver repeatedly fails to demonstrate he or she is ready to resume control of the car while the feature is on > Tesla has been pushing its driver-assist features, including Autopilot and what it calls “Full Self Driving,” which Tesla has insisted make driving safer than cars operated exclusively by humans. But NHTSA has been studying reports of accidents involving Autopilot and its Autosteer function for more than two years. > > The recall comes two days after a detailed investigation was published by the Washington Post that found at least eight serious accidents, including some fatalities, in which the the Autopilot feature should not have been engaged in the first place. > > Tesla’s owners manuals say: “Autosteer is intended for use only on highways and limited-access roads with a fully attentive driver.” But the company has pushed the idea that its driver assist features allow the cars to safely make most driving decisions even away from those roads. > > A NHTSA investigation, however, has found numerous accidents over the past several years that suggest that these features do not live up to their names of Autopilot and Full Self Driving. > > The safety regulator in its letter to Tesla said “in certain circumstances when Autosteer is engaged, the prominence and scope of the feature’s controls may not be sufficient to prevent driver misuse [of the feature.]” It said that when drivers are not fully engaged and ready to take control of the car “there may be an increased risk of a crash.” > > In addition to the software updates, Tesla will mail letters to car owners notifying them of the change. > A history of Autopilot issues > > This is not the first time that NHTSA has pushed Tesla to make changes to its Autopilot or Full Self Driving features after finding the features posed safety problems. > > In February, Tesla recalled all 363,000 US vehicles then on the road with its FSD feature after finding cars operating with the feature would violate traffic laws, including “traveling straight through an intersection while in a turn-only lane, entering a stop sign-controlled intersection without coming to a complete stop, or proceeding into an intersection during a steady yellow traffic signal without due caution.” > > And NHTSA and the National Transportation Safety Board have been investigating crashes involving Tesla vehicles using the various driver assist features, including a series of crashes into emergency vehicles on the scene of other accidents. > > Tesla is not the only automaker offering driver assist features marketed as “self-driving.” And it is not the only one to run into safety problems. Recently General Motors’ Cruise unit suspended its driverless taxi service nationwide after California authorities suspended its ability to operate the system there after an accident. > > But, because it markets the names Autopilot and Full Self Driving, Tesla has made a greater emphasis than competitors on self-driving. It charges buyers $6,000 for cars with what it calls “enhanced Autopilot.” and $12,000 for the FSD feature. > > Many who paid extra for those features have told CNN they think the features are not worth the extra money. But while the features have found support among other owners, the reports of serious accidents and deaths by police and safety regulators could hurt Tesla’s efforts to market the cars and their expensive features. > Autopilot’s importance to Tesla > > Tesla is already the most valuable automaker in the world, by far, despite having a fraction of the sales of many established automakers such as Toyota, Volkswagen, General Motors, Ford and Stellantis. > > Investors are betting on projections of future sales growth as well as the value of its software in making those stock valuations. CEO Elon Musk has said the company’s investment in artificial intelligence and its use in both self-driving vehicles as well as its plans for humanoid robots are a key to its current and future value. > > “In the long term, I think, has the potential to make Tesla the most valuable company in the world by far,” Musk said in October on a call with Wall Street analysts. “If you have fully autonomous cars at scale and fully autonomous humanoid robots that are truly useful, it’s not clear what the limit is.” > > Tesla’s stock fell slightly Wednesday. > > https://edition.cnn.com/2023/12/13/tech/tesla-recall-autopilot/index.html > > Qui il Safety Call Report: > https://static.nhtsa.gov/odi/rcl/2023/RCLRPT-23V838-8276.PDF > > ------------------------------ > > Subject: Digest Footer > > _______________________________________________ > nexa mailing list > nexa(a)server-nexa.polito.it > https://server-nexa.polito.it/cgi-bin/mailman/listinfo/nexa > > > ------------------------------ > > End of nexa Digest, Vol 176, Issue 18 > ************************************* ------------------------------------------------------------------------ the knowledge gap between rich and poor is widening witten & bainbridge, how to build a digital library ------------------------------------------------------------------------ Maurizio Lana Università del Piemonte Orientale Dipartimento di Studi Umanistici Piazza Roma 36 - 13100 Vercelli

1 0

A proposito di prodotti pericolosi e non funzionanti: Tesla recalls nearly all 2 million of its vehicles on US roads
by Daniela Tafani Dec. 14, 2023

Dec. 14, 2023

Tesla recalls nearly all 2 million of its vehicles on US roads Chris Isidore, CNN Tesla is recalling nearly all 2 million of its cars on US roads to limit the use of its Autopilot feature following a two-year probe by US safety regulators of roughly 1,000 crashes in which the feature was engaged. The limitations on Autopilot serve as a blow to Tesla’s efforts to market its vehicles to buyers willing to pay extra to have their cars do the driving for them. The National Highway Traffic Safety Administration said the Autopilot system can give drivers a false sense of security and be easily misused in certain dangerous situations when a Tesla’s technology may be unable to safely navigate the road. The over-the-air software update will give Tesla drivers more warnings when they are not paying attention to the road while the Autopilot’s “Autosteer” function is turned on. Those notifications will remind drivers to keep their hands on the wheel and pay attention to the road, according to a statement from NHTSA After the recall, Teslas with Autosteer turned on will more routinely check on the driver’s attention level – and may disengage the feature – when the software determines the driver isn’t paying attention, when the car is approaching traffic controls, or when it’s off the highway when Autosteer alone isn’t sufficient to drive the car. The recall was disclosed in a letter to Tesla posted by NHTSA, which said that Tesla had agreed to the software update starting on Tuesday that will limit the use of the Autosteer feature if a driver repeatedly fails to demonstrate he or she is ready to resume control of the car while the feature is on Tesla has been pushing its driver-assist features, including Autopilot and what it calls “Full Self Driving,” which Tesla has insisted make driving safer than cars operated exclusively by humans. But NHTSA has been studying reports of accidents involving Autopilot and its Autosteer function for more than two years. The recall comes two days after a detailed investigation was published by the Washington Post that found at least eight serious accidents, including some fatalities, in which the the Autopilot feature should not have been engaged in the first place. Tesla’s owners manuals say: “Autosteer is intended for use only on highways and limited-access roads with a fully attentive driver.” But the company has pushed the idea that its driver assist features allow the cars to safely make most driving decisions even away from those roads. A NHTSA investigation, however, has found numerous accidents over the past several years that suggest that these features do not live up to their names of Autopilot and Full Self Driving. The safety regulator in its letter to Tesla said “in certain circumstances when Autosteer is engaged, the prominence and scope of the feature’s controls may not be sufficient to prevent driver misuse [of the feature.]” It said that when drivers are not fully engaged and ready to take control of the car “there may be an increased risk of a crash.” In addition to the software updates, Tesla will mail letters to car owners notifying them of the change. A history of Autopilot issues This is not the first time that NHTSA has pushed Tesla to make changes to its Autopilot or Full Self Driving features after finding the features posed safety problems. In February, Tesla recalled all 363,000 US vehicles then on the road with its FSD feature after finding cars operating with the feature would violate traffic laws, including “traveling straight through an intersection while in a turn-only lane, entering a stop sign-controlled intersection without coming to a complete stop, or proceeding into an intersection during a steady yellow traffic signal without due caution.” And NHTSA and the National Transportation Safety Board have been investigating crashes involving Tesla vehicles using the various driver assist features, including a series of crashes into emergency vehicles on the scene of other accidents. Tesla is not the only automaker offering driver assist features marketed as “self-driving.” And it is not the only one to run into safety problems. Recently General Motors’ Cruise unit suspended its driverless taxi service nationwide after California authorities suspended its ability to operate the system there after an accident. But, because it markets the names Autopilot and Full Self Driving, Tesla has made a greater emphasis than competitors on self-driving. It charges buyers $6,000 for cars with what it calls “enhanced Autopilot.” and $12,000 for the FSD feature. Many who paid extra for those features have told CNN they think the features are not worth the extra money. But while the features have found support among other owners, the reports of serious accidents and deaths by police and safety regulators could hurt Tesla’s efforts to market the cars and their expensive features. Autopilot’s importance to Tesla Tesla is already the most valuable automaker in the world, by far, despite having a fraction of the sales of many established automakers such as Toyota, Volkswagen, General Motors, Ford and Stellantis. Investors are betting on projections of future sales growth as well as the value of its software in making those stock valuations. CEO Elon Musk has said the company’s investment in artificial intelligence and its use in both self-driving vehicles as well as its plans for humanoid robots are a key to its current and future value. “In the long term, I think, has the potential to make Tesla the most valuable company in the world by far,” Musk said in October on a call with Wall Street analysts. “If you have fully autonomous cars at scale and fully autonomous humanoid robots that are truly useful, it’s not clear what the limit is.” Tesla’s stock fell slightly Wednesday. https://edition.cnn.com/2023/12/13/tech/tesla-recall-autopilot/index.html Qui il Safety Call Report: https://static.nhtsa.gov/odi/rcl/2023/RCLRPT-23V838-8276.PDF

1 0

EU updates product liability regime to include software, Artificial Intelligence
by Daniela Tafani Dec. 14, 2023

Dec. 14, 2023

EU updates product liability regime to include software, Artificial Intelligence By Luca Bertuzzi | EU policymakers reached a political agreement on Thursday (14 December) on legislation to bring the bloc’s product liability regime in line with technological developments, notably covering digital products like software, which includes Artificial Intelligence. The Product Liability Directive (PLD) provides people who have suffered material damage from a defective product with the legal basis to sue the relevant economic operators and seek compensation. The product manufacturers will be liable for defectiveness resulting from a component under its control, which might be tangible, intangible, or a related service, like the traffic data of a navigation system. Defectiveness A product is deemed defective when it does not provide the safety a person is entitled to expect based on the reasonable foreseeable use, legal requirements, and the specific needs of the group of users for whom the product is intended. One of the elements considered when assessing defectiveness is the capacity of the product to continue to learn and acquire new features or knowledge; wording meant to cover Artificial Intelligence technology based on machine learning techniques. Damages Under the PLD, material damage includes death, personal injury, psychological harm, and destruction of property. At the same time, national liability regimes might still regulate compensation for non-material damages, such as those resulting from discrimination. The concept of damage also includes the loss or corruption of data that is not used exclusively for professional purposes. The European Parliament suggested introducing a threshold for the value of the data to qualify as a material loss, but this idea was rejected in the talks with the Council. Disclosure of evidence The PLD introduces the possibility for claimants who have presented sufficient evidence to request that a national court mandates the defendant disclose the evidence at its disposal as long as necessary and proportionate. Conversely, the Parliament proposed allowing the defendant to ask for the disclosure of evidence at the disposal of the claimant under the same conditions. The disclosing party might request the national court to take the necessary measures to protect the confidentiality of disclosed information including trade secrets. Burden of proof Claimants must prove the product’s defectiveness, the damage suffered, and the causal link between the two. However, under certain conditions, the defectiveness of the product will be assumed, and it will be on the defendant to disprove it. The conditions include cases where the defendant fails to disclose relevant evidence, the claimant demonstrates that the product does not comply with mandatory product safety requirements, or an obvious malfunction causes the damage. The defectiveness is also presumed in cases where the claimant faces excessive difficulties due to technical or scientific complexity to prove the defectiveness of the product or the causal link, or the claimant proves that it is likely that the product is defective or that there is a causal link. Open source software The Directive will not apply to free and open-source software developed or supplied outside a commercial activity. The liability rules apply when the software is supplied in exchange for a price or personal data used for anything other than improving the software’s security or compatibility. Compensation fund The European Parliament included the possibility of EU countries using existing or new national sectorial compensation schemes for victims of defective products who fail to obtain compensation because the economic operator is insolvent or no longer exists. The text specifies that the member states should preferably avoid funding the schemes with public money. Liability exemption The Parliament pushed a liability exemption for manufacturers of software components to a defective product that were micro or small enterprises when they placed that software in the market, provided that another economic operator is liable. However, this measure was left for contractual arrangements between the small company and the other party. Right of recourse Parliamentarians included the principle that whenever more than one economic operator is deemed liable for the same damage and is ordered to pay compensation, each one will have the right of recourse against the others. Risk development The Council added the possibility for EU countries to maintain existing measures in their legal systems that make economic operators liable even if they can prove that they could not know the product’s defectiveness based on the state of scientific and technical knowledge. Limitation periods Claimants will have three years to initiate proceedings following the damage. The right for an injured person to seek compensation under this directive expires after ten years, except for latent injuries that are slow to manifest, for which the timeframe is 25 years. Timeline The PLD will apply to all products placed on the EU market 24 months after it enters into force. EU countries will have until then to transpose the directive into national law. [Edited by Zoran Radosavljevic] L'articolo è in questo luogo occhiuto: https://www.euractiv.com/section/digital/news/eu-updates-product-liability-…

1 0

Dropbox spooks users by sending data to OpenAI for AI search features
by Daniela Tafani Dec. 13, 2023

Dec. 13, 2023

Dropbox spooks users by sending data to OpenAI for AI search features Feature turned on by default bothers users who like to keep their files private On Wednesday, news quickly spread on social media about a new enabled-by-default Dropbox setting that shares your Dropbox data with OpenAI for an experimental AI-powered search feature. Dropbox says that user data shared with third-party AI partners isn't used to train AI models and is deleted within 30 days. Even with assurances of data privacy laid out by Dropbox on an AI privacy FAQ page, the discovery that the setting had been enabled by default upset some Dropbox users. The setting was first noticed by writer Winifred Burton, who shared information about the Third-party AI setting through Bluesky on Tuesday, and frequent AI critic Karla Ortiz shared more information about it on X. Ortiz expressed worries that the data might be trained secretly without consent. In its FAQ, Dropbox contradicts this claim, saying, "We won’t let our third-party partners train their models on our user data without consent." Either way, communication about the change could have been clearer. AI researcher Simon Willison wrote, "Great example here of how careful companies need to be in clearly communicating what's going on with AI access to personal data." So why would Dropbox send user files to OpenAI anyway? In July, the company announced an AI-powered feature called Dash that allows AI models to perform universal searches across platforms like Google Workspace and Microsoft Outlook. According to the Dropbox privacy FAQ, the third-party AI opt-out setting is part of the "Dropbox AI alpha," which is a conversational interface for exploring file contents that involves chatting with a ChatGPT-style bot using an "Ask something about this file" feature. To make it work, an AI language model similar to the one that powers ChatGPT (like GPT-4) needs access to your files. According to the FAQ, the third-party AI toggle in your account settings is turned on by default if "you or your team" are participating in the Dropbox AI alpha. Still, multiple Ars Technica staff who had no knowledge of the Dropbox AI alpha found the setting enabled by default when they checked. Right now, the only third-party AI provider for Dropbox is OpenAI, writes Dropbox in the FAQ. "Open AI is an artificial intelligence research organization that develops cutting-edge language models and advanced AI technologies. Your data is never used to train their internal models, and is deleted from OpenAI’s servers within 30 days." It also says, "Only the content relevant to an explicit request or command is sent to our third-party AI partners to generate an answer, summary, or transcript." Disabling the feature is easy if you prefer not to share Dropbox data with OpenAI. Log into your Dropbox account on a desktop web browser, then click your profile photo > Settings > Third-party AI. This link may take you to that page more quickly. On that page, click the switch beside "Use artificial intelligence (AI) from third-party partners so you can work faster in Dropbox" to toggle it into the "Off" position. We reached out to Dropbox for comment, but it did not respond before this story was published. We will update this piece when we hear back from the company. https://arstechnica.com/information-technology/2023/12/dropbox-spooks-users…

1 0

Twenge, "Yes, it’s the phones (and social media)"
by J.C. DE MARTIN Dec. 13, 2023

Dec. 13, 2023

*Yes, it’s the phones (and social media) */How every other explanation – from lack of independence to academic pressure to the pandemic – falls short/ * *Jean M. Twenge Oct 18, 2023 * */https://jeanmtwenge.substack.com/p/yes-its-the-phones-and-social-media /* ------ Here are 13 Other Explanations For The Adolescent Mental Health Crisis. None of Them Work. */Only smartphones plus social media can explain the international collapse in the early 2010s. / Jean M. Twenge Oct 24, 2023 https://www.afterbabel.com/p/13-explanations-mental-health-crisis

3 2

Re: [nexa] AI systems? unlawful by default
by maurizio lana Dec. 13, 2023

Dec. 13, 2023

Il 13/12/23 14:05, Enrico Nardelli <nardelli(a)mat.uniroma2.it> ha scritto: > /This paper proposes a system of “unlawfulness by default” for AI systems, an ex-ante model where some AI developers have the burden of proof to demonstrate that their technology is not discriminatory, not manipulative, not unfair, not inaccurate, and not illegitimate in its legal bases and purposes/ > > https://brusselsprivacyhub.com/2022/06/01/brussels-privacy-hub-working-pape… > > Con un grazie a Daniela Tafani che me l'ha fatto scoprire in una sua recente presentazione questo paper del maggio 2022 è probabilmente una prima fase di riflessioni che hanno poi dato luogo a Malgieri, Gianclaudio, e Frank Pasquale. «Licensing High-Risk Artificial Intelligence: Toward Ex Ante Justification for a Disruptive Technology». /Computer Law & Security Review/ 52 (aprile 2024): 105899. https://doi.org/10.1016/j.clsr.2023.105899. anche questo, se non ricordo male, segnalato qui in lista da Daniela Tafani Maurizio PS: non è pignoleria documentale, ma il senso di ricostruire i percorsi degli sviluppi di pensiero, per sottrarsi almeno un po' all'inevitabile sensazione serendipica (almeno per me) che le cose interessanti si trovano qua e là basta che uno guardi con attenzione. che è pure vero, ma questo rende difficile cogliere delle linee di sviluppo > > Ciao, Enrico ------------------------------------------------------------------------ a questo punto devo fare una confessione: come il mio amico Erri De Luca, sono un europeista estremista. Questo significa che, per me, l’Europa unita è l’unica utopia politica ragionevole che noi europei abbiamo coniato. xavier cercas, inaugurazione del salone del libro, torino 2018 ------------------------------------------------------------------------ Maurizio Lana Università del Piemonte Orientale Dipartimento di Studi Umanistici Piazza Roma 36 - 13100 Vercelli

2 1

AI systems? unlawful by default
by Enrico Nardelli Dec. 13, 2023

Dec. 13, 2023

/This paper proposes a system of “unlawfulness by default” for AI systems, an ex-ante model where some AI developers have the burden of proof to demonstrate that their technology is not discriminatory, not manipulative, not unfair, not inaccurate, and not illegitimate in its legal bases and purposes/ https://brusselsprivacyhub.com/2022/06/01/brussels-privacy-hub-working-pape… Con un grazie a Daniela Tafani che me l'ha fatto scoprire in una sua recente presentazione Ciao, Enrico -- -- EN https://www.hoepli.it/libro/la-rivoluzione-informatica/9788896069516.html ====================================================== Prof. Enrico Nardelli Presidente di "Informatics Europe" Direttore del Laboratorio Nazionale "Informatica e Scuola" del CINI Dipartimento di Matematica - Università di Roma "Tor Vergata" Via della Ricerca Scientifica snc - 00133 Roma home page: https://www.mat.uniroma2.it/~nardelli blog: https://link-and-think.blogspot.it/ tel: +39 06 7259.4204 fax: +39 06 7259.4699 mobile: +39 335 590.2331 e-mail: nardelli(a)mat.uniroma2.it online meeting: https://blue.meet.garr.it/b/enr-y7f-t0q-ont ====================================================== --

1 0

Google’s Epic Legal Defeat Threatens $200 Billion App Store Industry
by Daniela Tafani Dec. 12, 2023

Dec. 12, 2023

Google’s legal defeat at the hands of Fortnite maker Epic Games Inc. threatens to roil an app store duopoly with Apple Inc. that generates close to $200 billion a year and dictates how billions of consumers use mobile devices. The loss — handed down by a San Francisco jury on Monday — is a blow to the two companies’ business model in apps, where they charge commissions of as much as 30% to software developers who typically have few other options. Google shares were down less than 1% in New York Tuesday morning at $132.62. Epic has spent years railing against the practice and got a federal jury to agree that Alphabet Inc.’s Google unit had acted unfairly as a monopoly. The case is likely to accelerate the weakening of app store rules, which have already come under fire from regulators and lawmakers around the world. “The dominoes are going to start falling here,” Tim Sweeney, chief executive officer of Epic, said in an interview after the verdict. “The end of 30% is in sight.” Though Apple won a similar case against Epic in 2021, that ruling was made by a single judge. The nature of the Google suit — where a jury sided unanimously with Epic — let actual consumers weigh in on the world of smartphone apps. In under four hours of deliberations, they found that Google had engaged in anticompetitive conduct, harmed Epic and illegally forced its own billing system on developers. There were other significant differences between the Apple and Google trials. During the current case, Epic highlighted agreements Google reached with top game developers, including Activision Blizzard Inc. and Nintendo Co., for smaller fees. “Revenue sharing deals among Google, smartphone makers and game developers came to light during the trial,” Justin Patterson and other analysts at KeyBanc Capital Markets wrote in a note to clients. “We believe this was a key difference between the cases that contributed to Apple’s victory and Google’s loss.” In Apple’s App Store, the same 30% take rate was applied evenly, while Google was attempting to steer traffic away from rival app stores on Android, analysts at Jefferies wrote in a note. “Meanwhile, Apple simply does not allow any rival app stores whatsoever.” The battle began in 2020, when Fortnite was kicked off the Apple and Google Play app stores because the game developer installed its own payment system. The idea was to bypass the up-to-30% revenue share that the two tech giants take from in-app purchases and subscriptions on their platforms. In response, Epic sued both companies. Google also has drawn criticism for making side deals with big developers like Spotify Technology SA where it offers lower commissions. In Monday’s decision, the jury found that Google shouldn’t require Android app developers to use its billing system for software sold through its store — and that it shouldn’t offer custom agreements to certain developers. “The immediate aftereffect is we will see a shift in the marketplace where big tech companies will have to make accommodations — whether it is more access, better terms, more options for developers — to stave off legal exposure,” said Paul Swanson, a partner at Holland & Hart who specializes in technology and antitrust law. Continua qui: https://www.bloomberg.com/news/articles/2023-12-12/google-s-epic-legal-defe…

1 0

MEMENTO | 167° Mercoledì di Nexa | 13 dicembre 2023, ore 17.00
by Valeria Bergantino Dec. 11, 2023

Dec. 11, 2023

Gentilissime, gentilissimi, Vi ricordiamo che *mercoledì 13 dicembre*, alle ore 17.00, si terrà il *167° Mercoledì di Nexa* con un incontro dal titolo /*"Internet fatta a pezzi. Sovranità digitale, nazionalismi e big tech"*/. Ospiti dell'incontro: *Vittorio Bertola* (Open-Xchange) e *Stefano Quintarelli *(UN/CEFACT)/*.*/ _L'incontro si terrà IN PRESENZA e ONLINE._ _SEDE FISICA_ dell'incontro: Centro Nexa su Internet e Società, Politecnico di Torino, Via Boggio 65/a, Torino (1° piano). Per accedere alla sala si raccomanda di suonare al citofono *Portineria* e di seguire le indicazioni segnalate lungo il percorso. QUI <https://nexa.polito.it/contatti> maggiori informazioni su come raggiungerci. _STANZA VIRTUALE_ dell'incontro: https://didattica.polito.it/VClass/NexaEvent Maggiori informazioni alla pagina: https://nexa.polito.it/mercoledi-167 Cordiali saluti, -- Valeria Bergantino Communication Officer Nexa Center for Internet & Society Politecnico di Torino - DAUIN Via Pier Carlo Boggio, 65/A - 10138 Torino web: https://nexa.polito.it/ <https://nexa.polito.it/> mail: valeria.bergantino(a)polito.it <mailto:valeria.bergantino@polito.it> tel: 3473443585

1 0

ChatGPT entusiasta dell'AI Act
by Guido Vetere Dec. 10, 2023

Dec. 10, 2023

GPT: The European Parliament and the Council have recently reached a provisional agreement on the Artificial Intelligence Act, a significant regulatory framework for the use of artificial intelligence (AI) in the European Union. The agreement aims to ensure that AI systems in the EU are safe and respect fundamental rights, while also fostering innovation and positioning Europe as a leader in the field. Key aspects of the Act include safeguards on general purpose AI, limitations on the use of biometric identification systems by law enforcement, bans on social scoring and AI used to manipulate or exploit vulnerabilities, and rights for consumers to lodge complaints and receive explanations. Specifically, the Act prohibits certain applications of AI that pose a threat to citizens' rights and democracy. These include biometric categorisation systems using sensitive characteristics, untargeted scraping of facial images, emotion recognition in workplaces and educational institutions, social scoring based on social behaviour, and AI systems that manipulate human behaviour or exploit vulnerabilities. For high-risk AI systems, the Act establishes clear obligations, including a mandatory fundamental rights impact assessment. This applies also to the insurance and banking sectors, and AI systems used in influencing elections and voter behaviour are classified as high-risk. The Act also sets out obligations for high-impact general-purpose AI systems, which include risk assessments, adversarial testing, incident reports, and transparency requirements. In terms of enforcement, non-compliance with the rules can result in fines ranging from 35 million euros or 7% of global turnover to 7.5 million euros or 1.5% of turnover, depending on the infringement and the size of the company. The agreed text will now need formal adoption by both the Parliament and Council to become EU law, with the Parliament’s Internal Market and Civil Liberties committees set to vote on the agreement in an upcoming meeting. The legislation is not expected to take effect until at least 2025. For more detailed information, you can refer to the articles from the European Parliament's news page [here](6) and from BNN [here](7). Me: yes, but consider the following passage: For high-impact GPAI models with systemic risk, Parliament negotiators managed to secure more stringent obligations. If these models meet certain criteria they will have to conduct model evaluations, assess and mitigate systemic risks, conduct adversarial testing, report to the Commission on serious incidents, ensure cybersecurity and report on their energy efficiency. MEPs also insisted that, until harmonised EU standards are published, GPAIs with systemic risk may rely on codes of practice to comply with the regulation. What 'codes of practice' means, in practice? Him: In the context of the European AI Act, "codes of practice" refer to a set of guidelines or standards that organizations can follow to ensure compliance with the regulation, particularly for general-purpose AI (GPAI) systems that pose systemic risks. These codes are intended to provide practical guidance on how to meet the Act's requirements, such as conducting model evaluations, assessing and mitigating systemic risks, performing adversarial testing, ensuring cybersecurity, and reporting on energy efficiency. Until harmonized EU standards are established, these codes of practice offer a framework for organizations to align their AI systems with the Act's objectives and principles. For more detailed information, you can refer to the European Parliament's website: [Artificial Intelligence Act: deal on comprehensive rules for trustworthy AI]( https://www.europarl.europa.eu/news/en/press-room/20231206IPR15699/artifici… ). Me: Please help me understand. What 'regulatory sandbox' could mean in the following passage? MEPs wanted to ensure that businesses, especially SMEs, can develop AI solutions without undue pressure from industry giants controlling the value chain. To this end, the agreement promotes so-called regulatory sandboxes and real-world-testing, established by national authorities to develop and train innovative AI before placement on the market Him: A "regulatory sandbox" in this context refers to a controlled environment where businesses, particularly small and medium-sized enterprises (SMEs), can develop and test innovative AI solutions under relaxed regulatory conditions. This setup allows for real-world testing and development without the full burden of compliance with extensive regulations that might normally apply. These sandboxes are established by national authorities to encourage innovation in AI while ensuring safety and compliance standards are met before these AI solutions are introduced to the broader market. For more detailed information, you can refer to the European Parliament's website: [Artificial Intelligence Act: deal on comprehensive rules for trustworthy AI]( https://www.europarl.europa.eu/news/en/press-room/20231206IPR15699/artifici… ) Me: Suppose I'm a SME and I'm in the process of developing an AI solution based on LLM. Why should I care about a “regulatory sandbox”? Him: As an SME developing an AI solution based on Large Language Models (LLM), a regulatory sandbox offers significant advantages. It allows you to test and refine your AI product in a real-world environment with reduced regulatory constraints. This means you can innovate and experiment more freely, accelerating development while ensuring compliance with essential safety and ethical standards. Essentially, a regulatory sandbox provides a safer, more flexible space for you to bring your AI innovations to market more efficiently and effectively. For more details, visit the European Parliament's website: [Artificial Intelligence Act: deal on comprehensive rules for trustworthy AI]( https://www.europarl.europa.eu/news/en/press-room/20231206IPR15699/artifici… ). Tutto a posto, buona domenica! Guido

1 0

Intelligenza artificiale: trovato accordo - via alla regolamentazione
by Mauro Alovisio Dec. 9, 2023

Dec. 9, 2023

Ciao segnalo news di interesse https://www.ilpost.it/2023/12/09/ai-act-intelligenza-artificiale-unione-eur… https://www.consilium.europa.eu/it/press/press-releases/2023/12/09/artifici… Buon sabato Mauro Alovisio

1 0

Attivissimo: Il video dimostrativo di Google Gemini è un falso
by Alberto Cammozzo Dec. 9, 2023

Dec. 9, 2023

<https://attivissimo.blogspot.com/2023/12/il-video-dimostrativo-di-google-ge…> Questo video promozionale di Google per presentare la propria nuova intelligenza artificiale, denominata /Gemini/, è un falso: Gemini non è affatto in grado di fare le cose mostrate qui. <https://www.youtube.com/embed/UIZAiXYceBI> Lo spiega in dettaglio Matteo Flora in questo video, supportato dalle dichiarazioni <https://developers.googleblog.com/2023/12/how-its-made-gemini-multimodal-pr…> fatte da Google stessa nei propri blog. <https://www.youtube.com/embed/rI0X6be5zC4> Specialmente per i chatbot, questa faccenda dell’IA comincia a puzzare sempre più di speculazione. /Fonti aggiuntive: Ars Technica <https://arstechnica.com/information-technology/2023/12/google-admits-it-fud…>, <https://arstechnica.com/information-technology/2023/12/google-admits-it-fud…> Gizmodo <https://gizmodo.com/google-gemini-ai-hands-on-demo-1851084170>. <https://gizmodo.com/google-gemini-ai-hands-on-demo-1851084170> /

2 1

Evaluating and Mitigating Discrimination in Language Model Decisions
by don Luca Peyron Dec. 8, 2023

Dec. 8, 2023

Buona lettura Dl https://www.anthropic.com/index/evaluating-and-mitigating-discrimination-in…

1 0