July 2020 - nexa - server-nexa.polito.it

Re: [nexa] face recognition a detrit

by Giovanni Biscuolo

Stefano Quintarelli <stefano(a)quintarelli.it> writes: > face recognition a Detroit > https://arstechnica.com/tech-policy/2020/06/detroit-police-chief-admits-fac… > sbaglia solo il 95-97% delle volte Interessante, diciamo che "fa curriculum" :-) ...però dobbiamo concedere loro che sbagliano così tanto SOLO perché - a quanto leggo - a Detroit sono molto attenti ad applicare la tecnica quasi solo nell'identificazione di persone afroamericane, nell'ambito di investigazioni sommarie alla carlona. Forse ci vorrebbero un po' più risorse per la ricerca sulla Scemenza Naturale invece che sulla Intelligenza Artificiale. [...] Saluti, Giovanni -- Giovanni Biscuolo

July 1, 2020

face recognition a detrit

by Stefano Quintarelli

face recognition a Detroit https://arstechnica.com/tech-policy/2020/06/detroit-police-chief-admits-fac… sbaglia solo il 95-97% delle volte ciao, s. -- reserve your meeting with me at https://cal.quintarelli.it

July 1, 2020

Webinar 'The Independent and Effective DPO: Legal and Policy Perspectives'

by Edoardo Celeste

Buonasera a tutti, di seguito vi prego di trovare i dettagli di un webinar che ho avuto recentemente modo di organizzare assieme a Future of Privacy Forum. Spero possa interessare alcuni di voi in lista. Qui <https://www.eventbrite.com/e/the-independent-and-effective-dpo-legal-and-po…> il link per la registrazione. Un caro saluto, Edoardo Celeste Future of Privacy Forum and Dublin City University invite you as we discuss: The Independent and Effective DPO: Legal & Policy Perspectives About this Event Join us Wednesday, July 8, 2020 at 3:00 PM CET (9:00 AM ET, 2.00 PM IST) until 4:30 PM CET for the webinar 'The independent and effective DPO: Legal and Policy Perspectives.' The webinar is jointly organized by Dublin City University (DCU) and Future of Privacy Forum (FPF) and designed to help policymakers, regulators, and their staff better understand the legal views concerning the position of the Data Protection Officer within an organization. Guest speakers include: MB Donnelly, Assistant Commissioner, Data Protection Commission Ireland Daniela Havlíková, Expert for the Legislation Unit, the Office for Personal Data Protection, Czech Republic Hielke Hijmans, Chairman of the Litigation chamber and member of the Board of Directors at Data Protection Authority Belgium Grégoire Delette, Legal advisor - DPOs dpt – Directorate of compliance, CNIL Marie Murphy, Fort Privacy and Association of Data Protection Officers Ireland Udo Oelen, Chief Privacy Officer at Dutch Railways, former Head of Private Sector at the Dutch Data Protection Authority Damien Kieran, Data Protection Officer at Twitter Mark Crosbie, Data Protection Officer at Dropbox Gabriela Zanfir-Fortuna, Senior Counsel, Future of Privacy Forum Rob van Eijk, Managing Director for Europe, Future of Privacy Forum Edoardo Celeste, Lecturer in Law, School of Law and Government, Dublin City University Please register in advance for this webinar, as space is limited. Registration will close at 5:00 PM ET on July 7th. Remote participation will be encouraged through live Q&A. DCU and FPF have partnered to collaborate on research projects, host joint conferences and workshops, and pursue applications for research opportunities together. To learn more about FPF in Europe, please visit https://fpf.org/eu <https://fpf.org/eu>. For information about DCU click here <https://www.dcu.ie/law_and_government/index.shtml>.

July 1, 2020

Facial recognition has a new foe: The world's biggest group of computing professionals

by Alberto Cammozzo

Il riconoscimento facciale farà la fine del poligrafo e del body scanner? Tecnologie ineffettive, inutili, in certi casi illegali, ma elementi redditizi del "security theater". <https://www.nbcnews.com/tech/security/tech-worker-group-calls-facial-recogn…> The world's largest association of computing professionals is calling for an "immediate suspension" of the private and governmental use of facial recognition technologies for "technical and ethical reasons." The Association for Computing Machinery, which says it has almost 100,000 student and professional members, said in a statement Tuesday that the surveillance technology was often biased and couldn't reliably used in a way that wouldn't adversely affect vulnerable populations. "The technology too often produces results demonstrating clear bias based on ethnic, racial, gender, and other human characteristics recognizable by computer systems," the group's U.S. Technology Policy Committee said in a statement. "Such bias and its effects are scientifically and socially unacceptable." The association acknowledged that facial recognition technology can be "benign or beneficial," but it said its use has "often compromised fundamental human and legal rights of individuals to privacy, employment, justice and personal liberty." It called on policymakers to support the suspension until legal standards for accuracy, fairness and accountability have been developed. Protests across the United States against police violence and racial profiling have galvanized researchers and civil liberties groups, who have been calling for more public scrutiny and strict controls of surveillance technologies, including facial recognition, which critics say aggravates human biases and infringes on people's constitutional freedoms. At the end of 2019, researchers for the National Institute of Standards and Technology, an agency of the Commerce Department, found that facial recognition algorithms falsely identified African American and Asian faces 10 to 100 times more often than Caucasian faces — a flaw that was illustrated by the wrongful arrest of Robert Williams, a Michigan man, this year. The Association for Computing Machinery follows a string of organizations and companies to issue warnings or bans on the technology. IBM, Amazon and Microsoft announced sweeping restrictions on their sales of facial recognition tools this month and called for them to be federally regulated.

July 1, 2020

Data, Compute, Labour | Ada Lovelace Institute

by J.C. DE MARTIN

https://www.adalovelaceinstitute.org/data-compute-labour/ Data, Compute, Labour30 June 2020 | Dr Nick Srnicek, King's College London The monopolisation of AI is not just – or even primarily – a data issue. Monopolisation is driven as much by the barriers to entry posed by fixed capital, and the ‘virtuous cycles’ that compute and labour are generating for the AI providers. (Sent from my wireless device; please excuse brevity and typos (if any))

July 1, 2020