July 2022 - nexa - server-nexa.polito.it

Articolo di Romano Prodi su politica industriale da Il Messaggero di ieri, 17 luglio 2022
by Luigi Scorca July 18, 2022

July 18, 2022

1 0

Facial Recognition Search Engine Pulls Up “Potentially Explicit” Photos of Kids
by Alberto Cammozzo July 17, 2022

July 17, 2022

<https://theintercept.com/2022/07/16/facial-recognition-search-children-phot…> Abusive parents searching for kids who have fled to shelters. Governments targeting the sons and daughters of political dissidents. Pedophiles stalking the victims they encounter in illicit child sexual abuse material. The online facial recognition search engine PimEyes allows anyone to search for images of children scraped from across the internet, raising a host of alarming possible uses, an Intercept investigation has found. Often called the Google of facial recognition, PimEyes search results include images that the site labels as “potentially explicit,” which could lead to further exploitation of children at a time when the dark web has sparked an explosion of images of abuse. “There are privacy issues raised by the use of facial recognition technology writ large,” said Jeramie Scott, director of the Surveillance Oversight Project at the Electronic Privacy Information Center. “But it’s particularly dangerous when we’re talking about children, when someone may use that to identify a child and to track them down.” Over the past few years, several child victim advocacy groups have pushed for police use of surveillance technologies to fight trafficking, arguing that facial recognition can help authorities locate victims. One child abuse prevention nonprofit, Ashton Kutcher and Demi Moore’s Thorn, has even developed its own facial recognition tool. But searches on PimEyes for 30 AI-generated children’s faces yielded dozens of pages of results, showing how easily those same tools can be turned against the people they’re designed to help. While The Intercept searched for fake faces due to privacy concerns, the results contained many images of actual children pulled from a wide range of sources, including charity groups and educational sites. PimEyes previously came under fire for including photos scraped from major social media platforms. It no longer includes those images in search results. Instead, searches churn up a welter of images that feel plucked from the depths of the internet. Some come from personal websites that parents created anonymously or semi-anonymously to feature photos of their children, likely not anticipating that they could one day be pulled up by strangers taking snapshots of kids on the street. One search for an AI-generated child turned up images of a real boy in Delaware, where a photographer had taken portraits of his family on a sunny spring day. When she posted the portraits in her online portfolio, the photographer omitted the boy’s name and other identifying details. But a determined person might theoretically be able to find such information. (The photographer did not respond to requests to comment for this article.) Another search turned up a girl displaying a craft project at an after-school program in Kyiv, Ukraine, in a photo taken just before the war. A second page on the same website showed the girl at home this spring; by then, Kyiv was under siege, the program had gone remote, and teachers were assigning kids craft projects to complete from their kitchen tables. A third search turned up a photo of a 14-year-old British boy that had been featured in a video about the U.K. educational system. The commentator gave the boy’s first name and details about the school he attended. Still another search turned up a photo of a toddler from an American home-schooling blog, where the girl’s mother had revealed her first name and, when the family was traveling, rough whereabouts. PimEyes is the brainchild of two Polish developers who created the site in 2017 on a whim. It reportedly passed through the hands of an anonymous owner who moved the headquarters to the Seychelles and then in December 2021 was purchased by Georgian international relations scholar Giorgi Gobronidze, who had met the site’s creators while lecturing in Poland. In a wide-ranging video interview that stretched to nearly two hours, Gobronidze offered a vague and sometimes contradictory account of the site’s privacy protections. He said that PimEyes was working to develop better safeguards for children, though he offered varying responses when asked what those might entail. “It’s a task that was given already to our technical group, and they have to bring me a solution,” he said. “I gave them several options.” At the same time, he dismissed the argument that parents who post anonymous photos of their children have any expectation of privacy. “Parents should be more responsible,” he said. “I have never posted a photo of my child on social media or on a public website.” “Designed for Stalkers” On its website, PimEyes maintains that people should only use the tool to search for their own faces, claiming that the service is “not intended for the surveillance of others and is not designed for that purpose.” But the company offers subscriptions that allow people to perform dozens of unique searches a day; the least expensive package, at $29.99 a month, offers 25 daily searches. People who shell out for the premium service can set alerts for up to 500 different images or combinations of images, so that they are notified when a particular face shows up on a new site. Gobronidze claimed that many of PimEyes’s subscribers are women and girls searching for revenge porn images of themselves, and that the site allows multiple searches so that such users can get more robust results. “With one photo, you can get one set of results, and with another photo you can get a totally different set of results, because the index combination is different on every photo,” he said. Sometimes, he added, people find new illicit images of themselves and need to set additional alerts to search for those images. He acknowledged that 500 unique alerts is a lot, though he said that, as of Thursday, 97.7 percent of PimEyes subscribers had a lighter account. Following criticism, the company pivoted to claiming that the search engine was a privacy tool. PimEyes’s previous owners marketed it as a way to pry into celebrities’ lives, the German digital rights site Netzpolitik reported in 2020. Following criticism, the company pivoted to claiming that the search engine was a privacy tool. Gobronidze said that fraught features were being overhauled under his ownership. “Previously, I can say that PimEyes was tailor-designed for stalkers, [in that] it used to crawl social media,” he said. “Once you dropped a photo, you could find the social media profiles for everyone. Now it is limited only to public searches.” But many people clearly do not see PimEyes as an aid to privacy. The site has already been used to identify adults in a wide variety of cases, from so-called sedition hunters working to find perpetrators after the January 6 insurrection, to users of the notorious site 4chan seeking to harass women. Nor do PimEyes’s marketing materials suggest much concern for privacy or ethics. In a version of the “people kill people” argument favored by the U.S. gun lobby, a blog post on the site blithely alludes to its many uses: “PimEyes just provides a tool, and the user is obliged to use the tool with responsibility. Everyone can buy a hammer, and everyone can either craft with this tool, or kill.” “These things should only be instrumentalized with the clear and knowledgeable consent of users,” said Daly Barnett, a staff technologist at the Electronic Frontier Foundation. “This is just another example of the large overarching problem within technology, surveillance-built or not. There isn’t privacy built from the get-go with it, and users have to opt out of having their privacy compromised.” “We Do Not Want to Be a Monster Machine” Alarmingly, search results for AI-generated kids also include images that PimEyes labels as “potentially explicit.” The backgrounds in the labeled images are blurred, and since clicking through to the source URLs could contribute to the exploitation of children, The Intercept could not confirm whether they are, in fact, explicit. Gobronidze said that the labels are assigned in part based on images’ source URLs, and that often the photos are harmless. When PimEyes representatives do run across child sexual abuse images, he said, representatives report it to law enforcement. But one example he gave shows how easily the site can be used to unearth abusive or illegal content. A 16-year-old girl had used her parents’ credit card to open an account, Gobronidze said. She soon found revenge porn videos that had been uploaded by an ex-boyfriend — images that likely fit the legal definition of child pornography. (He said PimEyes issued takedown notices for the websites and advised the girl to talk with authorities, her parents, and psychologists.) Gobronidze was vague on how he might limit abuse of children on the site. Subscribing requires a credit card, PayPal, or Amazon Pay account, and users upload their IDs only when asking PimEyes to perform takedown notices on their behalf. By design, he said, the search engine only seeks matches in photos and does not guess at age, gender, race, or ethnicity. “We do not want to be a monster machine,” he said, dubbing a more heavy-handed approach “Big Brother.” But at another point in the interview, he said he was planning to exclude images of children from search results. Still later, he said that his technical team was figuring out how to balance these two conflicting goals. PimEyes flags people who “systematically” use the engine to search for children’s faces, he said. Users who plug in one or two faces of children are typically assumed to be family members. If a PimEyes representative gets suspicious, he said, they might ask a subscriber for a document like a birth certificate that would prove that a user is a parent. When asked how a birth certificate would rule out abuse or stalking by noncustodial parents, Gobronidze said that PimEyes might instead request a signed form, similar to what parents and legal guardians provide in some countries when crossing borders with a child, to show they have any other parent’s consent. In a later email, he said that PimEyes had twice asked for “documents + verbal explanation” for people who uploaded images of children, and that the site had subsequently banned one of the accounts. “The fact that PimEyes doesn’t have safeguards in place for children and apparently is not sure how to provide safeguards for children only underlines the risks of this kind of facial recognition service,” said Scott, of EPIC. “Participating in public, whether online or offline, should not mean subjecting yourself to privacy-invasive services like PimEyes.” The inclusion of children’s faces in PimEyes search results underscores just how fraught the facial recognition landscape has become. The inclusion of children’s faces in PimEyes search results underscores just how fraught the facial recognition landscape has become. For years, victim advocacy groups have pushed for expanded use of the technology by law enforcement. The Kutcher-Moore nonprofit, Thorn, has developed a facial recognition tool called Spotlight that it provides to investigators working on sex trafficking cases, as well as to the National Center for Missing and Exploited Children. In a recent report, the center said that in 2021, Spotlight helped it identify over 400 missing children in online sex trafficking advertisements. Commercial providers of facial recognition have also gotten into trafficking prevention. The controversial facial recognition company Clearview AI sells its tools to police for identifying child victims. But those same tools can also be used to target the vulnerable. Clearview AI promoted the use of its database for child trafficking after being sued by the American Civil Liberties Union for endangering survivors of domestic violence and undocumented immigrants, among others. Prostasia Foundation, a child protection group that supports sex workers rights and internet freedom, contends that an earlier Thorn tool sometimes flagged images of adults, leading to the arrest of sex workers. This tension is even more extreme with PimEyes, which has virtually no guardrails and smashes long-standing expectations of privacy for both adults and children. Gobronidze said that PimEyes had talked to Thorn about using its tool Safer to detect child sexual abuse material using image hashing technology — a potentially odd relationship given that PimEyes makes images of children searchable to the general public, while Thorn aims to protect children from stalkers and abusers. “There has been one exploratory call between our Safer team and PimEyes to show how Safer helps platforms detect, report and remove CSAM,” a Thorn spokesperson said, using the acronym for child sexual abuse material. “No partnership materialized after that single call and they are not users of Safer or any tools built by Thorn.” When asked about concerns about its facial recognition tool, Thorn sent a statement through a spokesperson. “Spotlight is a highly targeted tool that was built specifically to identify child victims of sex trafficking and is only available to law enforcement officers who investigate child sex trafficking.” In the United States, PimEyes could run up against a 1998 law requiring the Federal Trade Commission to protect children’s online privacy. But so far, U.S. regulators have homed in on sites that store images or information, said Emma Llansó, director of the Free Expression Project at the Center for Democracy and Technology. PimEyes crawls images hosted on other sites. “PimEyes is just scraping whatever they can get their hands on on the web and isn’t making promises to users about what it will and won’t do with that data,” Llansó said. “So it’s something of a gray area.” Gobronidze is keenly aware of the distinction. “We don’t store any photos,” he claimed. “We don’t have any.” That is not entirely true. PimEyes’s privacy policy holds that for unregistered users — anyone who uses the site without a paid account — it retains facial images, along with the “fingerprint” of a face, for 48 hours and that data from the photos indexed in results is stored for two years. A sample PimEyes search showed thumbnail images of faces — photos returned in search queries that the site has edited to blur their backgrounds. A network traffic analysis showed that those photos are hosted on a PimEyes subdomain called “collectors.” In an email, Gobronidze said he had not previously heard or read about that subdomain and was “intrigued” to learn of it. He noted that he had forwarded the results of The Intercept’s analysis to PimEyes’s tech and data security units, adding that he could not “disclose [the] full technological cycle” because it is proprietary. Scott, of EPIC, would rather not wait around for courts and regulators to consider the storage question. “Congress needs to act to not only protect our children, but all of us from the dangers of facial recognition technology,” he said. “Services like this should be banned. That’s how you should regulate it.” -- PimEyes è interessante perché sfida le attuali regolazioni: non memorizza foto, ma è solo un crawler che indicizza le URL pubbliche alle quali trova i volti. E' quanto avevo in mente nel 2012 scrivendo "Do we need an open face recognition search engine?" <http://sedici.unlp.edu.ar/handle/10915/23870> L'unica differenza è che PimEyes è a pagamento ed è pensato per volumi di ricerche superiori a un solo volto, mentre io immaginavo un servizio pubblico eventualmente collegato a una procedura di take-down nel caso le immagini siano pubblicate illegittimamente.

1 0

China's Surveillance State Hits Rare Resistance From Its Own Subjects - Forbes India
by Alberto Cammozzo July 16, 2022

July 16, 2022

<https://www.forbesindia.com/article/news/chinas-surveillance-state-hits-rar…> Chinese artists have staged performances to highlight the ubiquity of surveillance cameras. Privacy activists have filed lawsuits against the collection of facial recognition data. Ordinary citizens and establishment intellectuals alike have pushed back against the abuse of COVID tracking apps by authorities to curb protests. Internet users have shared tips on how to evade digital monitoring. As China builds up its vast surveillance and security apparatus, it is running up against growing public unease about the lack of safeguards to prevent the theft or misuse of personal data. The ruling Communist Party is keenly aware of the cost to its credibility of any major security lapses: Last week, it moved systematically to squelch news about what was probably the largest known breach of a Chinese government computer system, involving the personal information of as many as 1 billion citizens. The breach dealt a blow to Beijing, exposing the risks of its expansive efforts to vacuum up enormous amounts of digital and biological information on the daily activities and social connections of its people from social media posts, biometric data, phone records and surveillance videos. The government says these efforts are necessary for public safety: to limit the spread of COVID, for instance, or to catch criminals. But its failure to protect the data exposes citizens to problems like fraud and extortion and threatens to erode people’s willingness to comply with surveillance. “You never know who is going to sell or leak your information,” said Jewel Liao, a Shanghai resident whose details were among those released in the leak. “It’s just a bit unusual to see that even the police are vulnerable too.” China, which has been racing to implement one of the world’s toughest data privacy regimes, frequently excoriates companies for mishandling data. But authorities rarely point fingers at the country’s other top collector of personal information: the government itself. Security researchers say the leaked database, apparently used by the police in Shanghai, had been left online and unsecured for months. It was exposed after an anonymous user posted in an online forum offering to sell the vast trove of data for 10 Bitcoin, or about $200,000. The New York Times confirmed parts of a sample of the database released by the anonymous user, who posted under the name ChinaDan. The government’s efforts to institute safeguards have lagged behind its own push to collect information. In recent years, The Times has reviewed other leaked databases used by the police in China that were left online with little to no protection; some contained facial recognition records and ID scans of people in a Muslim ethnic minority region. Now there are signs that people are growing wary of the government and public institutions, too, as they see how their own data is being used against them. Last month, a nationwide outcry erupted over the apparent abuse of COVID-19 tracking technology by local authorities. Protesters fighting to recover their savings from four rural banks in the central Chinese city of Zhengzhou found that the mobile apps used to identify and isolate people who might be spreading COVID-19 had turned from green — meaning safe — to red, a designation that would prevent them from moving freely. “There is no privacy in China,” said Silvia Si, 30, a protester whose health code had turned red. Authorities in Zhengzhou, under pressure to account for the episode, later punished five officials for changing the codes of more than 1,300 customers. Even when the COVID-19 tracking technologies are used for their stated purpose, more people seem willing to ask if the surveillance is excessive. On Wednesday, a blogger in Beijing posted on Weibo that he was refusing to wear an electronic bracelet to track his movements while in isolation, saying that the device was an “electronic shackle” and an infringement on his privacy. The post was liked around 60,000 times, and users flooded his post with responses. Many said it reminded them of the treatment of criminals; others called it a ploy to surreptitiously collect personal information. The post was later taken down by censors, the blogger said. In recent years, individuals have sought to draw attention to privacy concerns. In 2019, a law professor in Hangzhou, a prominent tech hub in eastern China, sued a local zoo for forcing him to submit facial recognition data to enter, the first such lawsuit in China. He won the case. Starting in late 2020, several Chinese cities began banning neighborhood committees from forcing residents to undergo biometric monitoring to enter their compounds. Around the same time, toilet paper dispensers using facial recognition were removed from public bathrooms in the southern Chinese city of Dongguan following public outrage. In online forums like Zhihu, a Quora-like platform, Chinese users trade advice on how to evade surveillance (tips include wearing hats and masks and pointing flashlights at security cameras). More than 60% of Chinese people say facial recognition technology has been abused, according to a study of more than 20,000 Chinese jointly conducted in late 2020 by a Chinese think tank and a government task force. More than 80% expressed concern about whether and how facial recognition data would be stored. “The rise of the public’s awareness of data privacy is an inevitable trend,” said Dragon Zheng, an artist based in the southern province of Guangxi whose practice explores the interaction of technology and governance. In 2016, Zheng installed security cameras inside a large exhibition hall, which streamed live footage to a monitoring room set up in the center of the hall. Visitors were invited to enter the room, where they could manipulate the cameras and experience what Zheng called the feeling of “monitoring and being monitored, controlling and being controlled.” Still, he emphasized that the risks and advantages of technology were not unique to China. “Technology is like Pandora’s box,” Zheng said. “Once it’s open, how it is used depends on whose hands it falls into.” Few Chinese citizens have publicly questioned the government about its collection of personal data. Part of that could be a result of the government’s thorough censorship and the threats to personal safety of criticizing the government. But many residents also see the handover of data as a necessary trade-off for security and convenience. “There’s always been this split identity when it comes to privacy awareness in China,” said Samm Sacks, a researcher on technology policy at Yale Law School and New America. “People are far more trusting overall in how government entities handle their personal information and far more suspicious about the corporate sector.”

1 0

Borse di partecipazione per studenti al Festival di Informatici Senza Frontiere (Rovereto, 20-22 Ottobre 2022)
by Patrignani Norberto July 15, 2022

July 15, 2022

Carissim*, segnalo e invito a far circolare il bando in allegato per 15 borse di partecipazione per altrettanti studenti e studentesse delle scuole superiori, studenti e studentesse universitarie e dottorandi/e, le borse coprono le spese di viaggio, vitto ed alloggio a Rovereto durante il periodo di svolgimento del Festival di Informatici Senza Frontiere (Rovereto, 20-22 ottobre 2022). Grazie! Norberto Le borse copriranno le spese di viaggio, vitto ed alloggio a Roveretodurante il periodo di svolgimento del Festival.

1 0

Google Analytics 4, GDPR... e la società cibernetica che vogliamo costruire.
by Giacomo Tesio July 14, 2022

July 14, 2022

Salve a tutti! Spero di far cosa gradita nel riportarvi alcuni passaggi di un interessante dibattito pubblico che sta animando gli addetti ai lavori ma apparentemente non riesce ad uscire dal proprio alveo. Il casus belli è la legittimità della nuova versione Google Analytics, ma il dibattito si sta rapidamente allargando alla legittimità di tutti i trasferimenti transfrontalieri che comportano gli stessi problemi: - il primo articolo è di Pietro Biase, presidente di Open Genova [1] ed è intitolato "Privacy, Google Analytics 4 è fumo negli occhi: ecco perché non risolve il problema" https://www.agendadigitale.eu/sicurezza/privacy/monitorapa-risponde-a-googl… - il secondo articolo è di Roberto Guiotto e Matteo Zambon, fondatori di Tag Manager Italia [2] ed è intitolato "Google Analytics 4, non è illegale: come usarlo bene ed evitare problemi col Gdpr" https://www.agendadigitale.eu/sicurezza/privacy/google-analytics-4-non-e-il… - l'ultimo opera mia e consiste in una risposta ufficiale di Monitora PA: "Google Analytics 4: Monitora PA risponde a Tag Manager Italia" https://monitora-pa.it/2022/07/14/GoogleAnalytics4_MonitoraPA_risponde_a_Ta… Il dibattito in corso è ovviamente molto più esteso e cerchiamo di tenerne traccia da mesi sulla home page di Monitora PA [3], ma questi tre articoli hanno il pregio di costituire una sorta di confronto dialettico e civile fra valori ed interessi antitetici. Inoltre in essi vi è assoluto accordo su un punto, espresso impeccabilmente da Tag Manager Italia: > è importante sottolineare che ciò che viene contestato a Google > Analytics, ossia la potenziale violazione alle condizioni di > legittimità del trasferimento di dati al di fuori dell’Europa, è – > con buona probabilità – una caratteristica comune a centinaia (o più > probabilmente migliaia) di software di utilizzo comune nell’ambito > del web marketing e del web hosting. > > Non è un mistero che il web marketing contemporaneo si basi in > larghissima parte su tecnologie e servizi ospitati su server > controllati da aziende USA (es. Facebook Ads, Google Ads, YouTube > Ads, etc.) > > Di conseguenza [...] è ragionevole ritenere > che le piattaforme pubblicitarie Ads utilizzate dalla stragrande > maggioranza dei web marketer, ma anche i software di email marketing, > i CRM, le piattaforme di Content Delivery Network, i web hosting, > etc. si trovino in una posizione simile rispetto a quella di Google > Analytics. > > [...] è un dato di fatto che i principi di funzionamento di tutte le > piattaforme sopra citate (nonché la nazionalità delle aziende che le > erogano) non sono particolarmente dissimili da quelli di Google > Analytics. Anche le conclusioni partono da osservazioni condivise: > Il 60% dei cittadini europei acquista online; il fatturato globale > dell’e-commerce in Europa vale 732 miliardi di dollari; la sola spesa > in digital advertising in Europa è stimata in 57,64 miliardi di > dollari nel 2022. Le conclusioni economiche che ne deriviamo sono però antitetiche: Tag Manager Italia conclude chiedendosi > In attesa di un accordo politico e giuridico che regolamenti la > questione, chi difende la produttività delle aziende del settore > digital e tutela il posto di lavoro di chi in queste aziende è > occupato? Monitora PA invece pone un diverso interrogativo > chi difende la produttività delle aziende italiane ed europee che > competono con Google & friends, senza appoggi politici [4], senza > sovvenzioni militari [5] e... rispettando pure la Legge? Spero che tutto questo possa essere di vostro interesse e che magari vogliate anche voi arricchire questo dibattito con le vostre prospettive. A presto! Giacomo [1] https://associazione.opengenova.org/ [2] https://www.tagmanageritalia.it/ [3] https://monitora-pa.it/ [4] https://www.key4biz.it/piccolo-e-importante-anche-per-il-cloud-il-gigantism… [5] https://www.nbcnews.com/tech/tech-news/thousands-contracts-highlight-quiet-…

7 10

"Contra Chrome": il ruolo di Chrome (il browser) nell'universo Google
by Damiano Verzulli July 14, 2022

July 14, 2022

Molti, in lista, sono certo ne siano gia' a conoscenza. Ma a qualcuno potrebbe tornare utile... Mi riferisco a Scott McCloud [1], fumettista americano, che nel 2008 si vide commissionata da Google la realizzazione di un fumetto [2] su quello che, ai tempi, sarebbe stato il "nuovo" browser che stava per essere rilasciato: Google Chrome. Lo stesso Scott, a 14 anni di distanza, ha sentito l'esigenza di "auto-commissionarsi" un aggiornamento del proprio fumetto dovuto allo snaturamento di quello che aveva raccontato 14 anni prima. L'update, dal titolo "Contra Chrome", è questo [3] A pochi mesi dall'uscita di "Contra Chrome", l'Associazione "Copernicani" ne ha derivato la versione "in italiano", pubblicata recentemente [4] Certamente chi segue questa ML è in grado di fruire della versione inglese. Tuttavia, per favorire la diffusione di "Contra Chrome" anche --e soprattutto-- ad un pubblico di non-addetti-ai-lavori, ritengo sia utile far riferimento a quella in Italiano. La traduzione è eccellente. I contenuti sono estremamente chiari e ben argomentati. È decisamente un testo da diffondere, soprattutto percheé espone molto chiaramente, e con parole e concetti estremamente "semplici", gran parte di quei temi "alti" che, spesso, si discutono su questa lista (privacy, capitalismo della sorveglianza, modelli di business delle big-tech, possibili impatti a livello democratico, etc.) Un caro saluto, DV P.S.: ne esistono anche le versioni PDF, comode per l'off-line: inglese [5] e italiano [6] [1] https://scottmccloud.com/7-consulting/index.html [2] https://www.google.com/googlebooks/chrome/big_00.html [3] https://contrachrome.com/ [4] https://copernicani.it/un-fumetto-ci-spiega-come-chrome-sia-diventato-una-m… [5] https://contrachrome.com/ContraChrome_en.pdf [6] https://copernicani.it/wp-content/uploads/2022/07/ContraChrome_IT_final.pdf -- Damiano Verzulli e-mail: damiano(a)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

1 0

What the Metaverse Is (Really) and Why We Need to Know About It
by don Luca Peyron July 13, 2022

July 13, 2022

Un buon pezzo da un punto di vista diverso dal solito Buona lettura dl https://www.liebertpub.com/doi/10.1089/cyber.2022.0124 _________________________ don Luca Peyron Direttore Ufficio per la Pastorale Universitaria Arcidiocesi di Torino www.universitari.to.it via XX settembre 83, Torino tel. 011 5156239

1 0

ANNUAL CONFERENCE of the Master in Intellectual Property | 15 luglio 2022, ore 9.15-13.00
by Nexa Media July 13, 2022

July 13, 2022

Gentilissimi, Vi invitiamo a partecipare alla *Conferenza Annuale del Master in Intellectual Property*: /*International Seminar on "Intellectual Property, Resilience and Recovery"*/, organizzato dal World Intellectual Property Organization, ITC ILO e Università di Torino, che si terrà *venerdì 15 luglio*, dalle ore 9.15 alle ore 13.00, presso il Centro Internazionale di Formazione dell'Organizzazione Internazionale del Lavoro, Viale Maestri del Lavoro, 10 - Torino. L'evento sarà inoltre fruibile on-line qui <https://itcilo-org.zoom.us/j/81866346384?pwd=a0loNXRlbnBNeCs5M2ZlRGh2QW94dz…>. Passcode: 12345 Partecipano al convegno, tra gli altri: *Prof. Alessandro Cogo*, Faculty Fellow del Centro Nexa, Università degli studi di Torino; *Prof. Maurizio Borghi*, Trustee del Centro Nexa, Università degli studi di Torino; *Prof. Marco Ricolfi*, Co-Fondatore e Co-Direttore del Centro Nexa, Università degli studi di Torino. Maggiori informazioni e il programma completo della conferenza sono disponibili alla pagina: https://nexa.polito.it/annual-conference-master-intellectual-property <https://nexa.polito.it/annual-conference-master-intellectual-property> <https://nexa.polito.it/annual-conference-master-intellectual-property> Si prega di confermare la propria partecipazione, sia online che in presenza, al seguente indirizzo email: llm_ip(a)itcilo.org Cordiali saluti, -- Anita Botta Communication Manager Nexa Center for Internet & Society Politecnico di Torino – DAUIN Corso Duca degli Abruzzi, 24 - 10129 Torino web: https://nexa.polito.it/ mail: anita.botta(a)polito.it tel: 011 090 7219

1 0

"Meet the Ex-CIA Agents Deciding Facebook’s Content Policy"
by J.C. DE MARTIN July 13, 2022

July 13, 2022

*Meet the Ex-CIA Agents Deciding Facebook’s Content Policy * Alan Macleod July 12th, 2022 It is an uncomfortable job for anyone trying to draw the line between “harmful content and protecting freedom of speech. It’s a balance”, Aaron says. In this official Facebook video <https://about.facebook.com/regulations/>, Aaron identifies himself as the manager of “the team that writes the rules for Facebook”, determining “what is acceptable and what is not.” Thus, he and his team effectively decide what content the platform’s 2.9 billion active users see and what they don’t see. Aaron is being interviewed in a bright warehouse-turned-studio. He is wearing a purple sweater and blue jeans. He comes across as a very likable, smiley person. It is not an easy job, of course, but someone has to make those calls. “Transparency is incredibly important in the work that I do,” he says. Aaron is CIA. Or at least he was until July 2019, when he left his job as a senior analytic manager at the agency to become senior product policy manager for misinformation at Meta, the company that owns Facebook, Instagram and WhatsApp. In his 15-year career, Aaron Berman rose to become a highly influential part of the CIA. For years, he prepared and edited the president of the United States’ daily brief, “wr[iting] and overs[eeing] intelligence analysis to enable the President and senior U.S. officials to make decisions on the most critical national security issues,” especially on “the impact of influence operations on social movements, security, and democracy,” his LinkedIn profile <https://www.mintpressnews.com/wp-content/uploads/2022/07/FireShot-Capture-0…> reads. None of this is mentioned in the Facebook video. [...] continua qui: https://www.mintpressnews.com/meet-ex-cia-agents-deciding-facebook-content-…

1 0

Inside a radical new project to democratize AI - A group of over 1,000 AI researchers has created a multilingual large language model bigger than GPT-3—and they’re giving it out for free
by Vetro' Antonio July 12, 2022

July 12, 2022

Segnalo il lancio del nuovo modello linguistico BLOOM che potrebbe stabilire nuovi standard di responsabilità nel settore dell’ IA: secondo quanto riportato dall’articolo del MIT review indicato sotto: è stato costruito da più di 1000 ricercatori, usando una grande varietà di sorgenti dati per le diverse lingue, condividendo i dettagli sulla raccolta dei dati, l'addestramento, il test, e l’uso del modello è disciplinato da una licenza di “AI responsabile” https://www.technologyreview.com/2022/07/12/1055817/inside-a-radical-new-pr… antonio

1 0