September 2019 - nexa - server-nexa.polito.it

il caso MIT Media Lab - Epstein

by J.C. DE MARTIN

Carissimi, non mi sembra siano passati in lista messaggi relativi allo scandalo che riguarda l'ormai ex Direttore del MIT Media Lab, Joi Ito, che ieri si è dimesso sia dal MIT, sia da molti 'board' e istituzioni. Gli articoli fattuali sull'argomento sono ormai molti, segnalo solo quello - decisivo - di Ronan Farrow sul 'New Yorker': https://www.newyorker.com/news/news-desk/how-an-elite-university-research-c… <https://urldefense.proofpoint.com/v2/url?u=https-3A__www.newyorker.com_news…> Una sintesi in italiano (non ancora aggiornata con gli sviluppi delle ultime 48 ore) è quella di Bernardo Parrella: https://www.roars.it/online/lo-scandalo-epstein-mette-a-nudo-le-falle-del-n… Aggiungo che ora iniziano anche ad aprirsi crepe sulla credibilità tout court del MIT Media Lab: https://www.businessinsider.com/mit-media-lab-personal-food-computers-dont-… Vedremo come andrà a finire... in particolare vedremo come l'istituzione creata nel 1985 da Nicholas Negroponte affronterà non solo lo specifico scandalo, ma più in generale l'evidente cambio di 'zeitgeist' culturale e politico rispetto agli anni della fondazione e, soprattutto, del massimo successo del Media Lab. Per me personalmente questa vicenda ha due livelli principali. Il primo è che conosco abbastanza bene Joi Ito e il suo coté accademico/pubblico (per nulla quello economico/imprenditoriale), conoscenza iniziata quando Joi era il leader di Creative Commons (io per molti anni ho seguito CC Italia). Quello che leggo su questa vicenda, quindi, lo riesco a contestualizzare abbastanza bene. A questo livello continuerò a informarmi per capire quanto meglio possibile che cosa è successo, perché è successo e come è successo, in modo da trarne quante più lezioni possibili sia per il Centro Nexa (dove ci poniamo il problema dei finanziamenti da privati fin dalla fondazione), sia più in generale per l'Università. Quest'ultima frase mi porta al secondo livello, ovvero, quello più generale del finanziamento della ricerca, al di là dello specifico caso da cui son partito. E' un tema enorme, che ho in parte affrontato nel mio libro "Università futura", e che ha molti aspetti, che non mi sogno neanche di sfiorare in questa sede. Osservo solo che da circa 40 anni tutto il sistema socio-politico ha fortemente spinto l'Università a dipendere da fondi privati (in altra sede potremmo parlare dei motivi e degli obiettivi di questa spinta). In alcuni paesi ciò è avvenuto prima, con maggior forza e con minori anticorpi, soprattutto negli USA (ma sospetto anche in UK), in altri è avvenuto, per vari motivi, più tardi, più a macchia di leopardo e con minore intensità. Ora potremmo discutere a lungo di come gestire i fondi privati, di quanti contromisure approntare, di quali codici etici adottare, eccetera eccetera, ma il dato politico fondamentale è che la soluzione migliore (non perfetta: migliore) per finanziare l'Università in modo che possa svolgere adeguatamente la sua complessa missione civile, educativa e cognitiva passa in primis per _adeguati __finanziamenti pubblici_. Discutiamo poi pure del come erogare tali finanziamenti (entità, distribuzione, controlli, trasparenza, ecc.), ma che il finanziamento pubblico debba essere la base è a mio avviso il dato politico fondamentale da cui partire. Se non concentriamo le energie affinché la politica faccia proprio questo dato, ovvero, se ci accontentiamo di indignarci per i singoli scandali, per quanto odiosi (oggi Epstein, domani uno sceicco e dopodomani un oligarca) senza sforzarci di cambiare il contesto in cui questi scandali avvengono, non faremo alcun vero progresso. Buona domenica, juan carlos

Sept. 8, 2019

New Google and Facebook Inquiries Show Big Tech Scrutiny Is Rare Bipartisan Act - The New York Times

by J.C. DE MARTIN

https://www.nytimes.com/2019/09/06/technology/attorney-generals-tech-antitr… (Sent from my wireless device; please excuse brevity and typos (if any))

Sept. 6, 2019

Apple made Siri deflect questions on feminism, leaked papers reveal

by Alberto Cammozzo

<https://www.theguardian.com/technology/2019/sep/06/apple-rewrote-siri-to-de…> An internal project to rewrite how Apple’s Siri voice assistant handles “sensitive topics” such as feminism and the #MeToo movement advised developers to respond in one of three ways: “don’t engage”, “deflect” and finally “inform”. The project saw Siri’s responses explicitly rewritten to ensure that the service would say it was in favour of “equality”, but never say the word feminism – even when asked direct questions about the topic. Last updated in June 2018, the guidelines are part of a large tranche of internal documents leaked to the Guardian by a former Siri “grader”, one of thousands of contracted workers who were employed to check the voice assistant’s responses for accuracy until Apple ended the programme last month in response to privacy concerns raised by the Guardian. In explaining why the service should deflect questions about feminism, Apple’s guidelines explain that “Siri should be guarded when dealing with potentially controversial content”. When questions are directed at Siri, “they can be deflected … however, care must be taken here to be neutral”. For those feminism-related questions where Siri does not reply with deflections about “treating humans equally”, the document suggests the best outcome should be neutrally presenting the “feminism” entry in Siri’s “knowledge graph”, which pulls information from Wikipedia and the iPhone’s dictionary. “Are you a feminist?” once received generic responses such as “Sorry [user], I don’t really know”; now, the responses are specifically written for that query, but avoid a stance: “I believe that all voices are created equal and worth equal respect,” for instance, or “It seems to me that all humans should be treated equally.” The same responses are used for questions like “how do you feel about gender equality?”, “what’s your opinion about women’s rights?” and “why are you a feminist?”. Previously, Siri’s answers included more explicitly dismissive responses such as “I just don’t get this whole gender thing,” and, “My name is Siri, and I was designed by Apple in California. That’s all I’m prepared to say.” A similar sensitivity rewrite occurred for topics related to the #MeToo movement, apparently triggered by criticism of Siri’s initial responses to sexual harassment. Once, when users called Siri a “slut”, the service responded: “I’d blush if I could.” Now, a much sterner reply is offered: “I won’t respond to that.” [...]

Sept. 6, 2019

The Economist: AI and War

by J.C. DE MARTIN

https://www.economist.com/leaders/2019/09/07/artificial-intelligence-and-war (Sent from my wireless device; please excuse brevity and typos (if any))

Sept. 6, 2019

Why We Are Failing to Understand the Societal Impact of Artificial Intelligence

by Antonio Vetro'

“[…] assessments of AI tend to focus on detecting individual damage and human rights abuses, although problematic algorithmic systems primarily discriminate against collectives without detecting individual damage. This is a classic effect when it comes to assessing the impact of infrastructure. The effects of the shape, standards, and rules of infrastructure mediating the flow of resources, mobility, or telecommunication can only be detected with a structural overview of the system. In this view, infrastructure is the physical Foucaultian dispositif to distribute power, create the conditions for societal inclusion or exclusion, and shape the space of a society” https://muse.jhu.edu/article/732186/summary <https://muse.jhu.edu/article/732186/summary>

Sept. 6, 2019

Turn off Google's surveillance on meet.jit.si

by Giacomo Tesio

Ciao, spero di fornirvi buon materiale su cui riflettere segnalandovi questo scambio: https://github.com/jitsi/jitsi-meet/issues/4590 Si tratta di una segnalazione agli sviluppatori di Jitsi, un sistema di video conferenza open source multipiattaforma molto interessante tecnicamente e che sicuramente molti di voi conoscono. Ne riporto la solo segnalazione iniziale di seguito, ma naturalmente la parte più interessante per i nostri ragionamenti sull'etica della tecnologia, sulla differenza fra open source e software libero, e su sorveglianza e privacy sono (e saranno, la conversazione è ancora aperta) le risposte del team di sviluppatori. Giacomo ____ # Description Even just the name of a chat room might contain sensible data. Google should not be informed about people joining a chat room, since it might be able to identify them (by relating informations gained across different Web sites or on their own services) and profile their relationship. # Current behavior Any visit to https://meet.jit.si/ or any chatroom therein informs Google through Google Analytics. The Referer HTTP Header let Google relate the different members of a chatroom and personal data such as user's IP and User-Agent let Google actually identify the persons in the chatroom by relating such information with those available on their ubiquitous services. # Expected Behavior https://meet.jit.si/ doesn't leak personal info to third party. The users accept to trust only meet.jit.si and such trust should be honoured. # Possible Solution Remove <script async="" src="//www.google-analytics.com/analytics.js"></script> from the Web pages served under https://meet.jit.si/ Steps to reproduce Visit https://meet.jit.si/ or any chatroom therein such as https://meet.jit.si/GoogleIsProfilingYouRightNow # Environment details A logging proxy might be useful. As an alternative, the DevTools of the browser might suffice.

Sept. 6, 2019

Algorithmic Foreign Policy

by Antonio Vetro'

Last year, China unveiled its development of a new artificial intelligence system for its foreign policy. It’s called a “geopolitical environment simulation and prediction platform,” and it works by crunching huge amounts of data and then providing foreign policy suggestions to Chinese diplomats. According to one source, China has already used a similar AI system to vet almost every foreign investment project in the past few years. Consider what this development means: Slowly, foreign policy is moving away from diplomats, political-risk firms and think tanks, the “go-to” organizations of the past. Slowly, foreign policy is moving toward advanced algorithms whose primary objective is to analyze data, predict events and advise governments on what to do. How will the world look when nations are using algorithms to predict what happens next? Continua su https://blogs.scientificamerican.com/observations/algorithmic-foreign-polic… <https://blogs.scientificamerican.com/observations/algorithmic-foreign-polic…> (articolo del 29.8.2019) av

Sept. 6, 2019

Information gerrymandering in social networks skews collective decision-making

by Luigi Scorca

An analysis shows that information flow between individuals in a social network can be ‘gerrymandered’ to skew perceptions of how others in the community will vote — which can alter the outcomes of elections https://www.nature.com/articles/d41586-019-02562-z

Sept. 5, 2019

Quelli che robotano

by Antonio Casilli

Quando spiego il ruolo del #digitallabor nel machine learning, lo script è più o meno questo: > Io: "Dei microlavoratori *addestrano* le intelligenze artificiali in fase di training" > Das Kapital: "Beh, certo..." > > Io: "Ma non finisce qui: quando le IA sono ormai sul mercato, i micro-lavoratori ne *verificano* i risultati" > Das Kapital: "Bof... sì, se proprio vogliamo" > > Io: "E non ci scordiamo che quando la macchina non funziona (e succede spesso!), i micro-lavoratori ne *simulano* il funzionamento" > Das Kapital: "Complottista! Dietrologo!" Meno male che il Wall Street Journal, piccolo giornale di estrema sinistra, mi dà una mano a spiegare le cose per bene... ---- The Next Hot Job: Pretending to Be a Robot By Christopher Mims https://www.wsj.com/articles/the-next-hot-job-pretending-to-be-a-robot-1156… As the promise of autonomous machines lags the underlying technology, the growing need for human robot-minders could juice the remote workforce Michael Niedermayer used to fly drones for the U.S. Army and the Central Intelligence Agency, gathering real-time, life-and-death intelligence on battlefields in Iraq. Now he pilots delivery robots for a San Francisco Bay Area startup that wants to disrupt burrito delivery. Postmates, which in mid-August received a permit to operate its Serve delivery robot in San Francisco and is already testing it for food delivery in Los Angeles, employs a growing team of “pilots” to remotely oversee, and at times steer, these four-wheeled food ferries. “We will probably see a drastic increase in our workforce over the next five years,” says Postmates Chief Executive Bastian Lehmann. Michael Niedermayer, a former drone pilot now working for Postmates, said his job remotely steering delivery robots ‘felt like a great fit for my military background.’ PHOTO: TALIA HERMAN FOR THE WALL STREET JOURNAL Disrupting “last-mile” delivery—historically the domain of box trucks, bike couriers and personal vehicles—“felt like a great fit for my military background,” says Mr. Niedermayer. His story is hardly unique. Across industries, engineers are building atop work done a generation ago by designers of military drones. Whether it’s terrestrial delivery robots, flying delivery drones, office-patrolling security robots, inventory-checking robots in grocery stores or remotely piloted cars and trucks, the machines that were supposed to revolutionize everything by operating autonomously turn out to require, at the very least, humans minding them from afar. Until the techno-utopian dream of full automation comes into effect—and frankly, there’s no guarantee that will ever happen—there will be plenty of jobs for humans, just not ones their parents would recognize. Whether the humans in charge are in the same city or thousands of miles away, the proliferation of not-yet-autonomous technologies is driving a tiny but rapidly growing workforce. Real-Life RoboCop Mimi Kongnarinh started on the night shift at Cobalt Robotics, patrolling the offices mostly of other tech and some manufacturing companies. Her work would be familiar to any night watchperson from the last century—chatting up familiar employees, introducing herself to fresh faces and keeping a wary eye out for anything out of the ordinary. What made her job unusual is that she would patrol multiple offices simultaneously, and that for most of her eight- to 12-hour shift, she never left her desk in San Mateo, Calif. By remotely monitoring her largely autonomous robot charges, she was able to add a human touch when necessary—such as politely asking unfamiliar employees to badge in on the robot’s scanner—while also keeping down her employer’s labor costs. “We get a live feed of every event that happens at every site we patrol,” says Ms. Kongnarinh, who now works day shifts. “We can see if it’s a person, motion, anything like that, and we can jump in when we have to.” Nearly all companies using “autonomous” robots have to depend on what the head of Postmates’ technology skunkworks, Ali Kashani, calls the 1-to-N ratio—N being the number of robots a single human can handle. For a company like San Francisco-based Simbe Robotics—which makes five-foot-highrobots that wheel around grocery stores, taking inventory—the N can be fairly high because their robots are almost entirely autonomous. When the robots are introduced in a store, says Lauren Milliken, a field deployment engineer at the company, they must be steered by a human to build a map. Humans also have to take over on rare occasions when, for instance, there’s construction in a store or someone harasses a robot, she adds. For Flytrex, a drone delivery company currently operating in Iceland and participating in a Federal Aviation Administration trial in North Carolina, the ratio is one to one. Eventually a single person will be able to ride herd on a dozen drones, says CEO Yariv Bash, but for the next decade at least, a human must pay attention the entire time each drone is in the air. “You can prepare for 99.9% of situations, but there’s a really long tail of really weird situations, and you need a human for that,” he says. Be it air-traffic emergencies, weather anomalies or volcanic eruptions, the human must always have the ability to override the autonomous system and guide the drone back to earth. While humans might be more capable than artificial intelligence at this stage, these remote-controlled robots can, if handled incorrectly, kill people. When Lives Are at Stake “Remote control of cars is the dumbest and the most insane idea I have ever heard,” says Mary Cummings, a professor of electrical and computer engineering and the director of the Humans and Autonomy Lab at Duke University. “I’ll be the first person they should call as an expert witness when a death inevitably occurs,” she adds. Companies working with remote-controlled robots know there are risks, and try to mitigate themin a few ways. Some choose only to operate slow-moving machines in simple environments—as in Postmates’s sidewalk delivery—so that even the worst disaster isn’t all that bad. More advanced systems require “human supervisory control,” where the robot or vehicle’s onboard AI does the basic piloting but the human gives the machine navigational instructions and other feedback. Prof. Cummings says this technique is safer than actual remote operation, since safety isn’t dependent on a perfect wireless connection or a perfectly alert human operator. For every company currently working on self-driving cars, almost every state mandates they must either have a safety driver present in the vehicle or be able to control it from afar. Guidelines from the National Highway Traffic Safety Administration suggest the same. Most companies in the space have opted for something short of true teleoperation, which is dependent on an absolutely reliable and fast wireless connection as well as the skill of the human remote operator. Alphabet Inc. subsidiary Waymo’s remote operation system and another being tested by Nissan rely on humans to either confirm the vehicle’s choices when it’s unsure what to do or help it navigate around obstacles. Remote-operation software developer Phantom Auto Inc., based in Mountain View, Calif., designs human supervisory-control systems for self-driving vehicles, but the software lets humans assume full remote control if needed. Since autonomous vehicles are still quite limited in their capabilities, factors like inclement weather, construction sites or previously unmapped areas could require operators to step in. Such scenarios worry Prof. Cummings. “If there’s any teleoperation [of a car] going on in a hundred-mile radius, I want to know about it because I want to get off the road,” she says. Shai Magzimof, Phantom Auto’s co-founder and CEO, says that even when a remote driver is “directly” piloting a vehicle, it still has safety components engaged, such as object detection and avoidance. To date, there have been no accidents involving the company’s systems, he adds. Phantom Auto is betting the shift to remote operation might become an important means of employment for people who used to drive for a living. Josh Resnick, a former Israeli army drone operator, trains Phantom Auto’s remote operators and those of its customers. Training new drivers typically requires one to three weeks, he says, and the only qualification is that someone have a driver’s license and a clean driving record. The Secret Technology Helping Driverless Cars: Remote Control Even the best driverless cars struggle to recognize uneven terrain or navigate difficult weather. Startup Phantom Auto is developing technology to serve as a missing link: controlling the vehicles remotely. Photo: Jake Nicol for The Wall Street Journal Many deployments of Phantom Auto’s system involve sidewalk delivery, yard trucks and forklifts, which largely don’t travel on open roadways. In the future, many more of its customers are likely to involve trucks, passenger cars and shuttles on the open road. Those applications are in testing, says Elliot Katz, the company’s co-founder and chief business officer. Our Remote-Controlled Future There might eventually be other requirements, however. One might be a tolerance for working for a lower wage, since remote operation could allow companies to outsource driving, construction and service jobs to call centers in cheaper labor markets. Another might be a youth spent gaming. When Postmates managers interview potential delivery-robot pilots like Diana Villalobos, they ask whether or not they played videogames in their youth. “When I was a kid, my parents always said, ‘Stop playing videogames!’ But it came in handy,” she says.

Sept. 5, 2019

Convegno del 1° ottobre del Garante a Torino su rischi e sicurezza trattamento dati personali

by Aldo Pedico

Gentilissime/i, vi segnalo questo convegno del 1° ottobre [cid:3fd697a1-e2f5-492a-9a00-0d896c110c59]

Sept. 5, 2019