- nexa - server-nexa.polito.it

Re: [nexa] Articolo What's wrong with "AI ethics" narratives

by Giuseppe Attardi

Grazie mille. Avevo intravisto il commento di Thomas Metzinger, ma non riuscivo a ritrovarlo. Sui social, il commento era stato criticato da Brian O’Sullivan, altro membro del comitato, dicendo che Metzinger ha firmato il documento senza obiezioni o senza proporre una posizione di minoranza. Ho sentito altre volte O’Sullivan prendere le difese della EC riguardo alle scelte in materia di AI. Temo che sia un atteggiamento comune tra gli esperti: meglio assecondare chi ci finanzia piuttosto che esprimere una posizione autonoma e critica. Purtroppo così facendo, stiamo infilandoci in un vicolo cieco. — Beppe > On 23 Oct 2022, at 15:01, Daniela Tafani <daniela.tafani(a)unipi.it> wrote: > > Grazie dell'osservazione. Ha ragione, sono stati anche gli esperti della Commissione. > Ci sono però testimonianze dirette del fatto che questi siano stati gentilmente invitati a farlo > da altri esperti, con incarichi direzionali e giunti lì attraverso sliding doors. > Mi sembra perciò un classico caso di cattura del regolatore, non trova? > E la cattura culturale converge verso lo stesso obiettivo. > > Riporto di seguito una parte della testimonianza di uno degli esperti europei incaricati (grassetto mio): > > As a member of the expert group, I am disappointed with the result that has now been presented. The guidelines are > lukewarm, short-sighted and deliberately vague. They ignore long-term risks, gloss over difficult problems (“explainability”) with > rhetoric, violate elementary principles of rationality and pretend to know things that nobody really knows. > Together with the excellent Berlin machine learning expert Urs Bergmann (Zalando), it was my task to develop, over many > months of discussion, the “Red Lines” – non-negotiable ethical principles determining what should not be done with AI in > Europe. The use of lethal autonomous weapon systems was an obvious item on our list, as was the AI-supported assessment > of citizens by the state (social scoring) and, in principle, the use of AIs that people can no longer understand and control. > I only realized that all this was not actually desired when our friendly Finnish HLEG President Pekka Ala-Pietilä (formerly > Nokia) asked me in a gentle voice whether we could remove the phrase “non-negotiable” from the document. In the next step, > many industry representatives and group members interested in a “positive vision” vehemently insisted that the phrase “Red > Lines” be removed entirely from the text – although it was precisely these red lines that were our mandate. The published > document no longer contains any talk of “Red Lines”; three were completely deleted and the rest were watered down. Instead > there is only talk of “critical concerns”. > > https://www.tagesspiegel.de/politik/ethics-washing-made-in-europe-5937028.h… <https://www.tagesspiegel.de/politik/ethics-washing-made-in-europe-5937028.h…> > > Analoghe pressioni hanno luogo nell'ambito privato, affinché i danni da decisioni algoritmiche siano presentati come problemi tecnici, > di moralizzazione dell'algoritmo. > Qui una testimonianza: > https://theintercept.com/2019/12/20/mit-ethical-ai-artificial-intelligence/ <https://theintercept.com/2019/12/20/mit-ethical-ai-artificial-intelligence/>. > > L'abuso dell'etica è una via di fuga dalla regolazione e, al momento, funziona benissimo, > quanto al raggiungimento dell'obiettivo prefissato. > > I risultati straordinari ci sono senz'altro, ma alcune decisioni oggi affidate a sistemi di ML ne sottintendono una visione magica: > animistica, antropomorfizzante, che mostra i risultati nascondendo lavoro e effetti collaterali, e affine all'astrologia. > > Grazie, buona domenica, > Daniela > > Da: nexa <nexa-bounces(a)server-nexa.polito.it> per conto di Giuseppe Attardi <attardi(a)di.unipi.it> > Inviato: domenica 23 ottobre 2022 12:57 > A: nexa(a)server-nexa.polito.it > Oggetto: Re: [nexa] Articolo What's wrong with "AI ethics" narratives > > "The question of algorithmic decisions is a political issue, which requires a political answer. Allowing algorithms to judge human beings is the same as deciding to generate exclusion and inequality on the basis of irrelevant factors for the profit of a few large private corporations." > > Sono d’accordo, come lo è anche Moshe Vardi, premio ACM Knuth 2022, che sostiene che affrontare il problema dell'uso dell'AI attraverso la lente dell’etica sia fuorviante, mentre invece il problema è inerentemente un problema politico (di policies). > > Però credo che a impostarlo in questo modo non siano state tanto le large corporations, a cui fa comunque comodo perché distoglie dai veri problemi di concentrazione e dominio economico, ma gli esperti selezionati dall’EC come consulenti e quelli che hanno portato alla stesura del demenziale European AI Act. > La EC, dopo un primo atteggiamento di interesse verso una strategia di sviluppo dell’AI in Europa, con la produzione del Coordinated Plan on AI del 2018, ha cambiato direzione spostando l'attenzione su questioni di regolamentazione. > L’European AI Act prevede la creazione di autorità di regolamentazione in ciascun paese dotate di adeguate risorse e un processo complicato di certificazione che lo stesso documento ritiene penalizzante per le SME. > A fronte di spese cospicue per realizzare queste strutture e i relativi “regulatory sandbox”, la UE investe cifre minuscole nella ricerca in AI. > > Quindi, invece di investire nel padroneggiare la tecnologia e in particolare saper decidere quali siano i suoi utilizzi migliori o peggiori, si investe nel tentativo di rallentarne l’adozione. > > L’AI sta facendo progressi straordinari (ne abbiamo accennato negli scambi sui LLM) e concentrandoci solo sulla regolamentazione, l’Europa sta perdendo l’occasione di avere un ruolo in una tecnologia dagli effetti dirompenti sulla società. > > — Beppe > >> On 14 Oct 2022, at 18:01, nexa-request(a)server-nexa.polito.it <mailto:nexa-request@server-nexa.polito.it> wrote: >> >> From: Daniela Tafani <daniela.tafani(a)unipi.it <mailto:daniela.tafani@unipi.it>> >> To: "nexa(a)server-nexa.polito.it <mailto:nexa@server-nexa.polito.it>" <nexa(a)server-nexa.polito.it <mailto:nexa@server-nexa.polito.it>> >> Subject: [nexa] Articolo What's wrong with "AI ethics" narratives >> Message-ID: <0aeb05e5f4724a5f91338c16bcb95b8c(a)unipi.it <mailto:0aeb05e5f4724a5f91338c16bcb95b8c@unipi.it>> >> Content-Type: text/plain; charset="windows-1252" >> >> Buongiorno. >> >> Propongo alla revisione paritaria aperta un mio articolo su What's wrong with "AI ethics" narratives: >> https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n… <https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n…> >> Daniela Tafani, What’s wrong with “AI ethics” narratives | Bollettino telematico di filosofia politica<https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n… <https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n…>> >> commentbfp.sp.unipi.it <http://commentbfp.sp.unipi.it/> >> PDF Abstract 1. Introduction 2. AI, magical thinking and imposture 3. A business model grounded on imposture: AI, cultural capture and regulatory capture 4. “AI ethics” narratives as mystification. The science fiction of AGI ethics… >> >> Grazie, saluti, >> Daniela Tafani

Oct. 24, 2022

Netflix: un algoritmo è per sempre

by don Luca Peyron

Interessanti sviluppi, buona lettura dl https://www.italian.tech/blog/upgrade/2022/10/24/news/netflix_un_algoritmo_… _________________________ don Luca Peyron Direttore Ufficio per la Pastorale Universitaria Arcidiocesi di Torino www.universitari.to.it via XX settembre 83, Torino tel. 011 5156239

Oct. 24, 2022

Cittadini ‘cibernetici’ contro capitalisti della sorveglianza

by de petra giulio

Sull’esperienza di MonitoraPa segnalo questa intervista a Giacomo Tesio. Giulio https://centroriformastato.it/cittadini-cibernetici-contro-capitalisti-dell…

Oct. 24, 2022

Connettività, non solo digital gap: la Gsma lancia l’allarme “digital usage”

by don Luca Peyron

Quanti divari che ci chiedono passi diversi Buona lettura dl https://www.corrierecomunicazioni.it/telco/digital-gap-gsma-nel-mondo-32-mi… _________________________ don Luca Peyron Direttore Ufficio per la Pastorale Universitaria Arcidiocesi di Torino www.universitari.to.it via XX settembre 83, Torino tel. 011 5156239

Oct. 24, 2022

Re: [nexa] Articolo What's wrong with "AI ethics" narratives

by Stefano Quintarelli

come sapete, io ero uno dei membri del gruppo di esperti. immagino non abbiate letto anche le policy recommendations. i deliverables furono 3, ma si parla solo di uno. chi segui' i lavori da vicino sa bene che solo una piccola parte delle nostre raccomandazioni fu adottata dalla politica. in merito alle "red lines", ricordo quel signore. in una delle occasioni in cui lo abbiamo visto (non molte, per la verita), all'inizio dei lavori, venne accompagnato da una troupe per farsi riprendere e andandosene da li' a poco, facendoci interrompere i lavori sul merito delle sue affermazioni, quello che ricordo e' che aveva proposto si' delle "red lines" riguardanti la superintelligenza e la creazione di macchine coscienti. avevamo prodotto due pagine al riguardo. poi ci siamo interrogati che farne: inserirle nel rapporto, avrebbe nobilitato un tema cui la quasi totalita' di noi dava lo stesso credito delle scie chimiche, toglierlo ci avrebbe esposti alla critica di non avere affrontato "questioni fondamentali per il futuro dell'umanita". con solo un paio di voci in dissenso, decidemmo di togliere le pagine ed inserire una nota a pie' pagina che diceva che ritenevamo certi temi fantascientifici, cosa che non piacque a quelli che su queste speculazioni ci stavano costruendo la propria immagine. lo ricordo bene perche' fui io che, un sabato mattina, feci un giro di consultazioni per cambiare le parole che dicevano che lo ritenevamo "science fiction" in "unrealistic". ciao, s. ps. con "sliding doors" si intende quando uno passa da regolatore a regolato. questo non e' il caso. il gruppo era composto da 52 persone (qualcuno e' venuto solo ad un paio delle molte decine di riunoini fatte). queste persone, me incluso, non erano retribuite ed appartenevano ad ogni tipo di organizzazione, da rappresentanti dei cosumatori, accademia, industria, politica, ecc. ci siamo auto-organizzati scegliendo come, quando lavorare ed attribuendoci i ruoli senza alcuna interferenza. e' tanto vero che persino l'incarico iniziale riguardava "trusted AI" e noi autonomamente ci siamo cambiati l'obiettivo parlando di "trustworthy AI". tanto vero che, nonostante la difesa militare sia un tema estraneo al perimetro della COmmissione UE e quindi al nostro incarico ed ambito dei documenti, ci siamo voluti comunque pronunciare sulle armi autonom, trascendendo i nostri compiti (sempre senza interferenza alcuna). queste cose le ho scritte all'epoca, ed in maggior dettaglio, quando il signore - out of the blue - si fece intervistare con quelle esternazioni ampiamente sconfessate. purtroppo non esistono depuratori per l'inquinamento informativo. ogni tanto riaffiora. ciao, s. On 23/10/22 15:01, Daniela Tafani wrote: > Grazie dell'osservazione. Ha ragione, sono stati anche gli esperti della Commissione. > > Ci sono però testimonianze dirette del fatto che questi siano stati gentilmente invitati a > farlo > > da altri esperti, con incarichi direzionali e giunti lì attraverso /sliding doors/. > > Mi sembra perciò un classico caso di cattura del regolatore, non trova? > > E la cattura culturale converge verso lo stesso obiettivo. > > > Riporto di seguito una parte della testimonianza di uno degli esperti europei incaricati > (grassetto mio): > > > /As a member of the expert group, I am disappointed with the result that has now been > presented. The guidelines are > lukewarm, short-sighted and deliberately vague. They ignore long-term risks, gloss over > difficult problems (“explainability”) with > rhetoric, violate elementary principles of rationality and pretend to know things that > nobody really knows. > Together with the excellent Berlin machine learning expert Urs Bergmann (Zalando), it was > my task to develop, over many > months of discussion, the “Red Lines” – non-negotiable ethical principles determining what > should not be done with AI in > Europe. The use of lethal autonomous weapon systems was an obvious item on our list, as > was the AI-supported assessment > of citizens by the state (social scoring) and, in principle, the use of AIs that people > can no longer understand and control. > I only realized that all this was not actually desired when *our friendly Finnish HLEG > President Pekka Ala-Pietilä (formerly > Nokia) asked me in a gentle voice whether we could remove the phrase “non-negotiable” from > the document*. In the next step, > *many industry representatives and group members interested in a “positive vision” > vehemently insisted that the phrase “Red > Lines” be removed entirely from the text – although it was precisely these red lines that > were our mandate. *The published > document no longer contains any talk of “Red Lines”; three were completely deleted and the > rest were watered down. Instead > there is only talk of “critical concerns”./ > > > https://www.tagesspiegel.de/politik/ethics-washing-made-in-europe-5937028.h… > <https://www.tagesspiegel.de/politik/ethics-washing-made-in-europe-5937028.h…> > > > Analoghe pressioni hanno luogo nell'ambito privato, affinché i danni da decisioni > algoritmiche siano presentati come problemi tecnici, > > di moralizzazione dell'algoritmo. > > Qui una testimonianza: > > https://theintercept.com/2019/12/20/mit-ethical-ai-artificial-intelligence/ > <https://theintercept.com/2019/12/20/mit-ethical-ai-artificial-intelligence/>. > > > L'abuso dell'etica è una via di fuga dalla regolazione e, al momento, funziona benissimo, > > quanto al raggiungimento dell'obiettivo prefissato. > > > I risultati straordinari ci sono senz'altro, ma alcune decisioni oggi affidate a sistemi > di ML ne sottintendono una visione magica: > animistica, antropomorfizzante, che mostra i risultati nascondendo lavoro e effetti > collaterali, e affine all'astrologia. > > > Grazie, buona domenica, > Daniela > > ------------------------------------------------------------------------------------------ > *Da:* nexa <nexa-bounces(a)server-nexa.polito.it> per conto di Giuseppe Attardi > <attardi(a)di.unipi.it> > *Inviato:* domenica 23 ottobre 2022 12:57 > *A:* nexa(a)server-nexa.polito.it > *Oggetto:* Re: [nexa] Articolo What's wrong with "AI ethics" narratives > "The question of algorithmic decisions is a political issue, which requires a political > answer. Allowing algorithms to judge human beings is the same as deciding to generate > exclusion and inequality on the basis of irrelevant factors for the profit of a few large > private corporations." > > Sono d’accordo, come lo è anche Moshe Vardi, premio ACM Knuth 2022, che sostiene che > affrontare il problema dell'uso dell'AI attraverso la lente dell’etica sia fuorviante, > mentre invece il problema è inerentemente un problema politico (di policies). > > Però credo che a impostarlo in questo modo non siano state tanto le large corporations, a > cui fa comunque comodo perché distoglie dai veri problemi di concentrazione e dominio > economico, ma gli esperti selezionati dall’EC come consulenti e quelli che hanno portato > alla stesura del demenziale European AI Act. > La EC, dopo un primo atteggiamento di interesse verso una strategia di sviluppo dell’AI in > Europa, con la produzione del Coordinated Plan on AI del 2018, ha cambiato direzione > spostando l'attenzione su questioni di regolamentazione. > L’European AI Act prevede la creazione di autorità di regolamentazione in ciascun paese > dotate di adeguate risorse e un processo complicato di certificazione che lo stesso > documento ritiene penalizzante per le SME. > A fronte di spese cospicue per realizzare queste strutture e i relativi “regulatory > sandbox”, la UE investe cifre minuscole nella ricerca in AI. > > Quindi, invece di investire nel padroneggiare la tecnologia e in particolare saper > decidere quali siano i suoi utilizzi migliori o peggiori, si investe nel tentativo di > rallentarne l’adozione. > > L’AI sta facendo progressi straordinari (ne abbiamo accennato negli scambi sui LLM) e > concentrandoci solo sulla regolamentazione, l’Europa sta perdendo l’occasione di avere un > ruolo in una tecnologia dagli effetti dirompenti sulla società. > > — Beppe > >> On 14 Oct 2022, at 18:01, nexa-request(a)server-nexa.polito.it >> <mailto:nexa-request@server-nexa.polito.it> wrote: >> >> From: Daniela Tafani <daniela.tafani(a)unipi.it <mailto:daniela.tafani@unipi.it>> >> To: "nexa(a)server-nexa.polito.it <mailto:nexa@server-nexa.polito.it>" >> <nexa(a)server-nexa.polito.it <mailto:nexa@server-nexa.polito.it>> >> Subject: [nexa] Articolo What's wrong with "AI ethics" narratives >> Message-ID: <0aeb05e5f4724a5f91338c16bcb95b8c(a)unipi.it >> <mailto:0aeb05e5f4724a5f91338c16bcb95b8c@unipi.it>> >> Content-Type: text/plain; charset="windows-1252" >> >> Buongiorno. >> >> Propongo alla revisione paritaria aperta un mio articolo su What's wrong with "AI >> ethics" narratives: >> https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n… >> <https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n…> >> Daniela Tafani, What’s wrong with “AI ethics” narratives | Bollettino telematico di >> filosofia >> politica<https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n… <https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n…>> >> commentbfp.sp.unipi.it <http://commentbfp.sp.unipi.it/> >> PDF Abstract 1. Introduction 2. AI, magical thinking and imposture 3. A business model >> grounded on imposture: AI, cultural capture and regulatory capture 4. “AI ethics” >> narratives as mystification. The science fiction of AGI ethics… >> >> Grazie, saluti, >> Daniela Tafani >> > > > _______________________________________________ > nexa mailing list > nexa(a)server-nexa.polito.it > https://server-nexa.polito.it/cgi-bin/mailman/listinfo/nexa

Oct. 23, 2022

Re: [nexa] Articolo What's wrong with "AI ethics" narratives

by Daniela Tafani

Grazie dell'osservazione. Ha ragione, sono stati anche gli esperti della Commissione. Ci sono però testimonianze dirette del fatto che questi siano stati gentilmente invitati a farlo da altri esperti, con incarichi direzionali e giunti lì attraverso sliding doors. Mi sembra perciò un classico caso di cattura del regolatore, non trova? E la cattura culturale converge verso lo stesso obiettivo. Riporto di seguito una parte della testimonianza di uno degli esperti europei incaricati (grassetto mio): As a member of the expert group, I am disappointed with the result that has now been presented. The guidelines are lukewarm, short-sighted and deliberately vague. They ignore long-term risks, gloss over difficult problems (“explainability”) with rhetoric, violate elementary principles of rationality and pretend to know things that nobody really knows. Together with the excellent Berlin machine learning expert Urs Bergmann (Zalando), it was my task to develop, over many months of discussion, the “Red Lines” – non-negotiable ethical principles determining what should not be done with AI in Europe. The use of lethal autonomous weapon systems was an obvious item on our list, as was the AI-supported assessment of citizens by the state (social scoring) and, in principle, the use of AIs that people can no longer understand and control. I only realized that all this was not actually desired when our friendly Finnish HLEG President Pekka Ala-Pietilä (formerly Nokia) asked me in a gentle voice whether we could remove the phrase “non-negotiable” from the document. In the next step, many industry representatives and group members interested in a “positive vision” vehemently insisted that the phrase “Red Lines” be removed entirely from the text – although it was precisely these red lines that were our mandate. The published document no longer contains any talk of “Red Lines”; three were completely deleted and the rest were watered down. Instead there is only talk of “critical concerns”. https://www.tagesspiegel.de/politik/ethics-washing-made-in-europe-5937028.h… Analoghe pressioni hanno luogo nell'ambito privato, affinché i danni da decisioni algoritmiche siano presentati come problemi tecnici, di moralizzazione dell'algoritmo. Qui una testimonianza: https://theintercept.com/2019/12/20/mit-ethical-ai-artificial-intelligence/. L'abuso dell'etica è una via di fuga dalla regolazione e, al momento, funziona benissimo, quanto al raggiungimento dell'obiettivo prefissato. I risultati straordinari ci sono senz'altro, ma alcune decisioni oggi affidate a sistemi di ML ne sottintendono una visione magica: animistica, antropomorfizzante, che mostra i risultati nascondendo lavoro e effetti collaterali, e affine all'astrologia. Grazie, buona domenica, Daniela ________________________________ Da: nexa <nexa-bounces(a)server-nexa.polito.it> per conto di Giuseppe Attardi <attardi(a)di.unipi.it> Inviato: domenica 23 ottobre 2022 12:57 A: nexa(a)server-nexa.polito.it Oggetto: Re: [nexa] Articolo What's wrong with "AI ethics" narratives "The question of algorithmic decisions is a political issue, which requires a political answer. Allowing algorithms to judge human beings is the same as deciding to generate exclusion and inequality on the basis of irrelevant factors for the profit of a few large private corporations." Sono d’accordo, come lo è anche Moshe Vardi, premio ACM Knuth 2022, che sostiene che affrontare il problema dell'uso dell'AI attraverso la lente dell’etica sia fuorviante, mentre invece il problema è inerentemente un problema politico (di policies). Però credo che a impostarlo in questo modo non siano state tanto le large corporations, a cui fa comunque comodo perché distoglie dai veri problemi di concentrazione e dominio economico, ma gli esperti selezionati dall’EC come consulenti e quelli che hanno portato alla stesura del demenziale European AI Act. La EC, dopo un primo atteggiamento di interesse verso una strategia di sviluppo dell’AI in Europa, con la produzione del Coordinated Plan on AI del 2018, ha cambiato direzione spostando l'attenzione su questioni di regolamentazione. L’European AI Act prevede la creazione di autorità di regolamentazione in ciascun paese dotate di adeguate risorse e un processo complicato di certificazione che lo stesso documento ritiene penalizzante per le SME. A fronte di spese cospicue per realizzare queste strutture e i relativi “regulatory sandbox”, la UE investe cifre minuscole nella ricerca in AI. Quindi, invece di investire nel padroneggiare la tecnologia e in particolare saper decidere quali siano i suoi utilizzi migliori o peggiori, si investe nel tentativo di rallentarne l’adozione. L’AI sta facendo progressi straordinari (ne abbiamo accennato negli scambi sui LLM) e concentrandoci solo sulla regolamentazione, l’Europa sta perdendo l’occasione di avere un ruolo in una tecnologia dagli effetti dirompenti sulla società. — Beppe On 14 Oct 2022, at 18:01, nexa-request(a)server-nexa.polito.it<mailto:nexa-request@server-nexa.polito.it> wrote: From: Daniela Tafani <daniela.tafani(a)unipi.it<mailto:daniela.tafani@unipi.it>> To: "nexa(a)server-nexa.polito.it<mailto:nexa@server-nexa.polito.it>" <nexa(a)server-nexa.polito.it<mailto:nexa@server-nexa.polito.it>> Subject: [nexa] Articolo What's wrong with "AI ethics" narratives Message-ID: <0aeb05e5f4724a5f91338c16bcb95b8c(a)unipi.it<mailto:0aeb05e5f4724a5f91338c16bcb95b8c@unipi.it>> Content-Type: text/plain; charset="windows-1252" Buongiorno. Propongo alla revisione paritaria aperta un mio articolo su What's wrong with "AI ethics" narratives: https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n… Daniela Tafani, What’s wrong with “AI ethics” narratives | Bollettino telematico di filosofia politica<https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n…> commentbfp.sp.unipi.it<http://commentbfp.sp.unipi.it/> PDF Abstract 1. Introduction 2. AI, magical thinking and imposture 3. A business model grounded on imposture: AI, cultural capture and regulatory capture 4. “AI ethics” narratives as mystification. The science fiction of AGI ethics… Grazie, saluti, Daniela Tafani

Oct. 23, 2022

Ginevra Sanvitale. Technopolitical Resonance: Emotions, Computers and Socialism in Cold War Italy (1965-1990)

by Federico Leva (Nemo)

Segnalo dissertazione recentemente difesa all'università di Eindhoven, su argomenti cari a questa lista. https://pure.tue.nl/ws/portalfiles/portal/217797547/20221019_Sanvitale_hf.p… Può interessare anche l'uso di archivi digitali come l'archivio Primo Moroni e Grafton9. Federico == Abstract == Emotions and politics are often considered an hindrance to technological development. Even more, technology is promised to simplify political processes, and to overcome emotional distress. Yet, after a century of these promises, the contemporary world looks more politically complex and emotionally demanding than ever. What if we change perspective, by looking at our tech-centered world from other dimensions of human experience? In this PhD dissertation, I show that “emotions,” “technology,” and “politics” are always profoundly interrelated, by presenting an in depth historical investigation of the role played by emotions in the re-politicization and de- politicization of computer technologies in Cold War Italy (1965-1990). I developed a novel methodological and conceptual approach centered on the notion of “Technopolitical Resonance,” to understand how emotions contributed to make specific technopolitical configurations more or less popular through history. This approach provides an actor-centered framework to investigate emotions’ significance in the History of Technology, currently lacking in the field. It is based on literature from the History and Anthropology of Emotions, stressing the epistemic and performative significance of emotions. The dissertation is centered on the reproduction and the rejection of a technopolitical configuration which I call “the Black Box Entanglement”. This configuration, I claim, relied on the “fear of falling behind” in the Cold War to promote computer use, and their design as “black boxes,” that users could not study nor modify. The dissertation critically analyzes the diffusion of black-boxed computers as a de-politicizing design choice, because the design process includes only a limited number of actors, namely the engineers and software developers, thus reducing the space for democratic participation. The dissertation also offers a critical perspective on “fear of falling behind” as a de-politicizing discourse on the societal significance of computers, because it flattens the political debate favoring a phenomenological approach (how can technology solve our problem? -Because it certainly will) over a dialectical one (why will technology -and not something else- solve our problem?). Several re-politicizing counter-narratives are also analyzed, based on different emotions (i.e. hope, anger, pride) and different political visions on the societal significance of computers and their design. == Preface (and acknowledgments) == How I learned to stop worrying, and love the history of technology I am at home in Bologna, working on my doctoral project about the history of computing. My partner, a software developer, is playing a part in unfolding this history from the other room. Our house is full of computer parts, computer books, and conversations about computers. What a modern couple we are, how fit for the Computer Age! But, every now and then, our day is interrupted by a computerized voice, which neither of us has programmed. “Linea 37” it says. It comes from an automatic speaker, announcing this bus’s arrival at the stop down in the street. Bus 37 is not like any other bus in Bologna. It is a memento, a window into a past which I have not lived, but can’t be forgotten. On August 2, 1980, bus 37 was temporarily converted into a hearse. At 10:25 that morning, a bomb exploded in Bologna’s central train station, killing 85 people and wounding hundreds. Bus 37 was used to transport the victims’ bodies to the morgue. Far from being an isolated episode, the Bologna Massacre was the most recent event in a decade-long period of bombings and other violent acts perpetrated by Neo-fascist groups, with the more or less tacit approval of the Italian secret services and armed forces. This was the so-called “Strategy of Tension,” aimed at destabilizing Italian public morale in order to shift the government to the farright. Today, most historians agree that such a plan had no chance of success, and this was already clear to many people at the time. However, the deaths caused by these attacks were very real, and disturbing events undeniably took place during the period known as the “Italian first republic” (1948-1994). Bus 37’s arrival has interrupted my work, so I decide to take a break and read something about the present. “Facebook broke democracy!” claim news outlets reporting on the Cambridge Analytica scandal, or some other wrongdoing by Mark Zuckerberg’s company. When 10 years ago the “Arab Spring” prompted an opposite claim, that social media could “make” democracy, a terrible delusion was around the corner. The Cambridge Analytica scandal and the Arab Spring delusion have indeed something in common: people lost sight of extremely complex societal and political processes in favor of a technology-centered, de-politicized vision. But, if it was true that a website could “break” or “make” a democracy, then perhaps our democracies were not so strong in the first place. This would make it even more imperative to address such events from a political, not technological perspective. What if it is never technology, but always politics that makes or breaks democracies? This dissertation is grounded in the idea that we, European scholars and citizens, 1 need to repoliticize public debates on technology, otherwise we will never be able to mend our democracies— which, if not broken, are certainly crooked. The word “politics” has acquired a negative connotation nowadays. You should not “discuss politics” at dinner parties, and take care not to look “too political” at work. But this word has been in our vocabularies for centuries, and the concept it describes has existed even longer. The fact that we don’t talk about politics, won’t make it disappear: only make it more difficult to understand how it works. And this is a problem for democracy, because if political processes are not based on transparency, and accountability, and participation, then perhaps we are not in a functioning democracy. Even in settings devoted to political decision making, technology-related choices seem to lack a coherent and explicit political rationale. European Union policymakers dream of “technological sovereignty,”2 but they also produce a “Digital Economy and Society Index” 3 which generically drives greater use of digital technologies, regardless of their provenance and software license. Things are no better when we look at grassroots political movements: recent years have seen the emergence of a transnational environmentalist movement, famously exemplified by the Fridays For Future activists, and a feminist “fourth wave” has apparently begun. However, technology-critical movements today are the Cinderella of grassroots politics, and the “fourth feminist wave” is one example: in the 1990s, the combination of feminism and technology meant cyborgs and selfmanaged servers, whereas now it means #hashtags on corporate-managed social media. Is this a feminist Arab Spring, or will the change last longer in this case? And I am still into “third wave” feminism: what does this make me? Am I already a #boomer, 4 in my early 30s? I start wondering, anxiously. But bus 37 is here again, shifting my thoughts back to the past. Had the Bologna massacre happened today, there would be hundreds of pictures and videos (and hashtags) about the event. The trials might have been over in a few years, with all the additional evidence available on Instagram. Perhaps my fourth wave feminist sisters are right, I am a #boomer. Maybe I should learn to stop worrying, and start loving Mark Zuckerberg. No, I should not. Because if there is something that all generations of feminists have in common is a unique emotional attitude, described by Sara Ahmed in her “Killjoy Manifesto.” 5 And this attitude is not exclusive to feminists: every individual wishing to engage critically with societal and political issues must be prepared to cause some degree of joy-killing. During my research, I encountered many killjoys. They critically examined “this circularity of illusions-delusions which follows each technological cycle, probably from the wheel to the steam machine, electricity, and automation.” 6 Killjoys who challenged the idea that a computer could “make” or “break” democracy, and therefore debated how to use this technology without falling for yet another “depressingly uninspiring”7 utopian (or dystopian) plan. Killjoys who knew very well that the Computer Age was the same “Age” when bombs exploded inside train stations, and bus 37 became a hearse. At times, I even felt overwhelmed by my sources: historical actors knew much more than me, about both politics and technology. Despite being so knowledgeable, they failed to produce a long-lasting repoliticization of computer debates: what made me think I could be more successful? While entangled in this unsolvable question, I also learned one certainty, as the Killjoy Manifesto’s 5th principle states: “I am not willing to get over histories that are not over.” There is a political and emotional history of the Computer Age which is certainly not over. It is a fragmented history, because the actors who made it were at times in conflict with each other. But it is also a contiguous history: these actors often had the same feelings and the same thoughts about computers’ political significance. It is a Resonant history, as I call it in this dissertation. Recomposing this history, looking at emotions as a shared space of understanding, is what I can add to the knowledgeable debates and analysis of the past. No, Facebook did not break democracy, and it won’t fix it. Forgetting is what broke democracy. And politics is what can fix it. Today we have many expectations about digital technologies, but sometimes we find comfort in these expectations to avoid tackling difficult and uncomfortable political issues. The road to the re-politicization of computer debates seems long and impervious, however we already have the intellectual and practical tools to aid us with this process. This is why I learned to stop worrying, and start loving the history of technology. *** Whereas learning how to love the history of technology was a very enjoyable process, I often wondered whether the history of technology would love me back, and this has been a frequent source of methodological anxieties and theoretical conundrums. I started this PhD project as a trained anthropologist. This proved to be an asset because I was drawn to actors and sources that are currently under-represented in the history of computing. But I frequently asked myself how the personal engagement required in anthropology could be reconciled with the distancing from historical actors often required by historiography. I am thus very grateful for the support and the encouragement I have received from my PhD promotor Erik van der Vleuten, my supervisor and initiator of the “Fearful technologies” project Karena Kalmbach, and my co-supervisor, Andreas Spahn. They provided stimulating questions and insights, while leaving me free to explore my own path and interests. I could not have asked for better mentors. Doctoral committee members Paul Edwards, Anna Guagnini, Ruth Oldenziel, and Valérie Schafer provided constructive and critical comments which greatly improved this manuscript. My first encounters with the research field happened during my years as a master student at the University of Bologna: I will always be grateful to Anna Guagnini and Giuliano Pancaldi, for introducing me to the fascinating world of the History of Science and Technology. The History Lab at Eindhoven University of Technology did the rest, providing engaging conversations and perspectives. I would like to thank Ruth Oldenziel, Mila Davis, Frank Veraart, Harry Lintsen, Eric Berkers, Jonas van der Straeten and Jan Korsten for the History Lab reading seminars (as well as the coffee breaks), and for their insights in studying, researching, and teaching the History of Technology. Through the Eindhoven History Lab I had the opportunity to discuss my work with colleagues who offered valuable comments on my drafts and research plans: Dick van Lente, Peter Norton, Arwen Mohun, Frank Schipper. A special mention to my PhD colleagues in the History Lab, Patrick Bek and Henk-Jan Dekker, with whom I shared the incredible adventure that is achieving a PhD (and during a pandemic!). Patrick’s intellectual insights and emotional support have been particularly important in the final months, when it seemed the project would never end. During my PhD project, I also had the opportunity to attend workshops, summer schools, and conferences. Three of these were particularly important for shaping this dissertation and I am indebted to the organizers and participants. I am particularly grateful to Martina Hessler and Bettina Hitzler for the workshop “The Multifacted Relationship between Fear and Technology,” held at the Max Planck Institute for Human Development in October 2018, and the resulting publication; to Valérie Schafer and Paul Edwards for the insights gained at the Tensions of Europe summer school “Towards Digital Science and Technology Studies: Challenges and Opportunities,” held at the Centre for Contemporary and Digital History (C²DH) at the University of Luxembourg in June 2019; to Aristotle Tympas for the Tensions of Europe workshop “Computing, Artificial Intelligence, Big Data, Algorithms, Internet of Things, Social Media, Automation, Robotics and Cybernetics: Historical and STS Perspectives from Mediterranean/Southern/Southeastern Europe,” held online in June 2021, as well as our conversations on the history of computing. I would also like to thank Anna Åberg, Andreas Marklund, Anique Hommels (and again Karena Kalmbach) for the panels on technology and crises we organized during Tensions of Europe conferences. Not only historians helped to shape my work. The Technology, Innovation and Society group at Eindhoven University of Technology provided a welcoming and stimulating environment during my PhD studies. The Thursday seminars offered very valuable interdisciplinary exchanges, as did the many informal gatherings with PhD and postdoc colleagues at TU/e. I am especially indebted to Ankit, Darja, Edgar, Matthew, Minha, Tanja and our visiting colleagues from abroad, Aske and Michal. My pre-doctoral years have been equally important for this dissertation. Professor Davide Domenici was the first to show me there was an incredibly fascinating world beyond the histoire événementielle I had been studying before taking his classes. This is the first dissertation I have written without his supervision. Other fundamental learning came from classes by Ivo Quaranta (Cultural and Medical Anthropology), Luca Jourdan (Social and Political Anthropology), Cristiana Natali and Gateano Mangiameli (Research Methodology). Furthermore, I would like to express my deep gratitude to the people and institutions who provided oral accounts and source materials for this dissertation. Paola Manacorda, Fiorella De Cindio, Filippo Demonte, IBM labor unionists Francesco Fiaccadori, Alfio Riboni, Giovanni Talpone, Valeria Bernardi and Renato Pomari, all graciously agreed to be interviewed, sharing precious information on Italy’s history of computing. I am grateful to the many archivists and cultural institutions that provided materials: staff at Archivio Gramsci Emilia-Romagna; Archivio RSU IBM; Centro di documentazione dei movimenti “F. Lorusso – C. Giuliani”; Centro Studi Libertari – Archivio Giuseppe Pinelli; Archivio Grafton9; Biblioteca Libertaria Armando Borghi; Biblioteca “Elio Xerri” – Circolo Anarchico “C. Berneri.” Deserving a special mention is Nicola, at Centro Lorusso – Giuliani, whose captivating enthusiasm and commitment made it very difficult to move on to a different archive. A researcher’s life is not only about doing research. Writing about your research is an equally fundamental task. It would be much more difficult for you to understand this manuscript if it wasn’t for the amazing editing work done by Val Kidd. I am also indebted to Anne Schuler, who introduced me to English academic writing. TIS secretariat support was fundamental to avoid succumbing to the intricacies of university regulations: thank you Letty Calame, Sonia Parker, and Iris Houx for your support and your patience. There is one emotion that most researchers share: a love of knowledge. I had the privilege to meet many people who encouraged and nurtured my love of knowledge. I would like to acknowledge, and honor some of those who saw the beginning of this research project but not its end. Professor Maria Maddalena Mené introduced me to Natalia Ginzburg and to cultural geography (I now realize this was the first step toward my anthropology degree). Professor Lucio Carugno taught me how to find the beauty and relevance of literary texts from the past. My uncle, professor Mario Bressan, did not succeed in turning my love for science into a chemistry degree, but he might have found some resonance in this PhD dissertation and I would have loved to discuss it with him. My grandmother, Gloria Papa, an expert in literature and an excellent writer: she would have been a great PhD candidate, if she had had the opportunity. My mother, Fabrizia Arduini, who always found time to take me to the public library as a kid, and loved this research project. Last but not least, I am deeply indebted to my family and friends, for their enduring patience, love, and encouragement. My family gave me great support over these years, especially my father, Alberto Sanvitale, my aunts, Alessandra and Sylvia Arduini, my cousin Elena Caputi, my “fairy godmothers” Annamaria Properzi, Laura Florani and Stefania De Carlo. Gaia, Giacomo, Giorgia, Margherita, Marta: you are “my people,” as Dr. Cristina Yang would say. Thank you for being such amazing human beings and friends. Barbara, Cristina, Noemi: our conversations on technology, politics, and feminism nurtured my intellectual life, as much as our friendship nurtured my days. Andrea, Mirna, Mauro, Nicola (and the other players mentioned here): thank you for our amazing Dungeons&Dragons adventures. With you I learned how it feels to have a class enemy and what can be achieved with the right party. About my real-life political education, there are too many people I should be thanking: from Pavia to Bologna, from the Wikiverse to the Fediverse, I have learned and grown a lot thanks to you—and hopefully helped to bring about some societal change in the process. Finally, my partner Cek, software developer, but also my sci-fi literature advisor and technical consultant. We went from a long-distance relationship to being stuck in the same apartment together during lockdown. I could imagine no better person by my side in these beautiful but also difficult years. Thank you for enduring my 24/7 listening to Italian political songs or Blondie, and for making sure I always had my morning coffee.

Oct. 23, 2022

Re: [nexa] The US supreme court case that could bring the tech giants to their knees | John Naughton | The Guardian

by Federico Leva (Nemo)

Ci andrei cauto. https://www.techdirt.com/2022/06/07/academic-paper-shows-how-badly-the-main… Federico

Oct. 23, 2022

Re: [nexa] Articolo What's wrong with "AI ethics" narratives

by Giuseppe Attardi

"The question of algorithmic decisions is a political issue, which requires a political answer. Allowing algorithms to judge human beings is the same as deciding to generate exclusion and inequality on the basis of irrelevant factors for the profit of a few large private corporations." Sono d’accordo, come lo è anche Moshe Vardi, premio ACM Knuth 2022, che sostiene che affrontare il problema dell'uso dell'AI attraverso la lente dell’etica sia fuorviante, mentre invece il problema è inerentemente un problema politico (di policies). Però credo che a impostarlo in questo modo non siano state tanto le large corporations, a cui fa comunque comodo perché distoglie dai veri problemi di concentrazione e dominio economico, ma gli esperti selezionati dall’EC come consulenti e quelli che hanno portato alla stesura del demenziale European AI Act. La EC, dopo un primo atteggiamento di interesse verso una strategia di sviluppo dell’AI in Europa, con la produzione del Coordinated Plan on AI del 2018, ha cambiato direzione spostando l'attenzione su questioni di regolamentazione. L’European AI Act prevede la creazione di autorità di regolamentazione in ciascun paese dotate di adeguate risorse e un processo complicato di certificazione che lo stesso documento ritiene penalizzante per le SME. A fronte di spese cospicue per realizzare queste strutture e i relativi “regulatory sandbox”, la UE investe cifre minuscole nella ricerca in AI. Quindi, invece di investire nel padroneggiare la tecnologia e in particolare saper decidere quali siano i suoi utilizzi migliori o peggiori, si investe nel tentativo di rallentarne l’adozione. L’AI sta facendo progressi straordinari (ne abbiamo accennato negli scambi sui LLM) e concentrandoci solo sulla regolamentazione, l’Europa sta perdendo l’occasione di avere un ruolo in una tecnologia dagli effetti dirompenti sulla società. — Beppe > On 14 Oct 2022, at 18:01, nexa-request(a)server-nexa.polito.it wrote: > > From: Daniela Tafani <daniela.tafani(a)unipi.it <mailto:daniela.tafani@unipi.it>> > To: "nexa(a)server-nexa.polito.it <mailto:nexa@server-nexa.polito.it>" <nexa(a)server-nexa.polito.it <mailto:nexa@server-nexa.polito.it>> > Subject: [nexa] Articolo What's wrong with "AI ethics" narratives > Message-ID: <0aeb05e5f4724a5f91338c16bcb95b8c(a)unipi.it <mailto:0aeb05e5f4724a5f91338c16bcb95b8c@unipi.it>> > Content-Type: text/plain; charset="windows-1252" > > Buongiorno. > > Propongo alla revisione paritaria aperta un mio articolo su What's wrong with "AI ethics" narratives: > https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n… <https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n…> > Daniela Tafani, What’s wrong with “AI ethics” narratives | Bollettino telematico di filosofia politica<https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n… <https://commentbfp.sp.unipi.it/daniela-tafani-what-s-wrong-with-ai-ethics-n…>> > commentbfp.sp.unipi.it <http://commentbfp.sp.unipi.it/> > PDF Abstract 1. Introduction 2. AI, magical thinking and imposture 3. A business model grounded on imposture: AI, cultural capture and regulatory capture 4. “AI ethics” narratives as mystification. The science fiction of AGI ethics… > > Grazie, saluti, > Daniela Tafani >

Oct. 23, 2022

The US supreme court case that could bring the tech giants to their knees | John Naughton | The Guardian

by Alberto Cammozzo

<https://www.theguardian.com/commentisfree/2022/oct/22/the-us-supreme-court-…> Two weeks ago, the US supreme court decided that it would hear Gonzalez v Google, a landmark case that is giving certain social-media moguls sleepless nights for the very good reason that it could blow a large hole in their fabulously lucrative business models. Since this might be good news for democracy, it’s also a reason for the rest of us to sit up and pay attention. First, some background. In 1996, two US lawmakers, Representative Chris Cox from California and Senator Ron Wyden from Oregon, inserted a clause into the sprawling telecommunications bill that was then on its way through Congress. The clause eventually became section 230 of the Communications Decency Act and read: “No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.” The motives of the two politicians were honourable: they had seen how providers of early web-hosting services had been held liable for damage caused by content posted by users over whom they had no control. It’s worth remembering that those were early days for the internet and Cox and Wyden feared that if lawyers had henceforth to crawl over everything hosted on the medium, then the growth of a powerful new technology would be crippled more or less from birth. And in that sense they were right. What they couldn’t have foreseen, though, was that section 230 would turn into a get-out-of-jail card for some of the most profitable companies on the planet – such as Google, Facebook and Twitter, which built platforms enabling their users to publish anything and everything without the owners incurring legal liability for it. So far-reaching was the Cox-Wyden clause that a law professor eventually wrote a whole book about it, The Twenty-Six Words That Created the Internet. A bit hyperbolic, perhaps, but you get the idea. Now spool forward to November 2015 when Nohemi Gonzalez, a young American studying in Paris, was gunned down in a restaurant by the Islamic State terrorists who murdered 129 other people that night. Her family sued Google, arguing that its YouTube subsidiary had used algorithms to push IS videos to impressionable viewers, using the information that the company had collected about them. Their petition seeking a supreme court review argues that “videos that users viewed on YouTube were the central manner in which IS enlisted support and recruits from areas outside the portions of Syria and Iraq which it controlled”. The key thing about the Gonzalez suit, though, is not that YouTube should not be hosting IS videos (section 230 allows that) but that its machine-learning “recommendation” algorithms, which may push other, perhaps more radicalising, videos, renders it liable for the resulting damage. Or, to put it crudely, while YouTube may have legal protection for hosting whatever its users post on it, it does not – and should not – have protection for an algorithm that determines what they should view next. This is dynamite for the social-media platforms because recommendation engines are the key to their prosperity. They are the power tools that increase the user “engagement” – keeping people on the platform to leave the digital trails (viewing, sharing, liking, retweeting, purchasing, etc) – that enable the companies to continually refine user profiles for targeted advertising. And make unconscionable profits from doing so. If the supreme court were to decide that these engines did not enjoy section 230 protection, then social media firms would suddenly find the world a much colder place. And stock-market analysts might be changing their advice to clients from “hold” to “sell”. Legal scholars have been arguing for decades that section 230 needs revision. Freedom of speech fanatics see it as a keystone of liberty, as the “kill switch” of the web. Donald Trump made threatening noises about it. Tech critics (such as this columnist) regard it as an enabler of corporate hypocrisy and irresponsibility. However you look at it, though, it’s more than half a century since it became law, which is about 350 years in internet time. Having such a statute to regulate the contemporary networked world seems a bit like having a man with a red flag walking in front of a driverless car. (Though, come to think of it, that might not be such a bad idea.) Versions of the question posed by the Gonzalez suit – whether section 230 immunises internet platforms when they make targeted recommendations of content posted by other users – have been put to US courts over the last few years. To date, five courts of appeals judges have concluded that the section does provide such immunity. Three appeals judges have ruled that it does not, while one other has concluded only that legal precedent precludes liability for recommendation engines. There’s no legal consensus here, in other words. It’s high time that the supreme court decided the matter. After all, isn’t that what its there for? .

Oct. 22, 2022