Caroline Lequesne

Caroline Lequesne

@clequesne.bsky.social

PhD, MCF HDR - Ass. Prof Université Côte d’Azur Law, vulnerability and algorithms. French academic & (enthusiast) global scholar.

607 followers250 following183 posts

Caroline Lequesne reposted

L’omnibus numérique veut affaiblir le RGPD au profit des constructeurs d’IA. Il leur permettrait notamment d’utiliser les données des travailleurs sans leur consentement. C’est une attaque contre nos droits et c’est aujourd’hui qu’il faut les défendre ! euobserver.com/239590/eu-da...

euobserver.comEU data chief adds voice to unions’ fears of ‘carte blanche’ use of workers’ data for AIEU data protection chief Wojciech Wiewiórowski has joined trade unions and privacy activists in their concerns over the ‘digital omnibus’, which includes amendments to EU privacy rules that would allo...

Caroline Lequesne reposted

Caroline Lequesne reposted

Caroline Lequesne reposted

This is insanity. Irish Presidency's draft text of amendments to the GDPR allows training & operation of any AI without safeguards & tough legal basis (under "legitimate interest" a company claims what it is doing is OK. Up to you to find out and challenge in court). This is a free pass for US Tech.

Violation de données personnelles par un agent IA: 1ere notification à l’autorité 🇪🇸 de protection des données. Un changement de nature des cyberattaques: l’enjeu n’est plus seulement l’assistance à un humain mais la possibilité de la perpétuer de façon autonome. www.reuters.com/business/spa...

reuters.comSpanish data watchdog publicises first AI agent-linked data breach reportSpain's data protection watchdog said it has received the first reported notification of a personal data breach allegedly carried out by an artificial intelligence agent, a case that suggests autonomo...

IA, fin du monde et valorisation des géants. Combien vaut une entreprise qui pourrait éradiquer l’humanité ? Quelques réflexions au départ de la question impertinente soulevée cette semaine par Nicolas Chapuis dans Le Monde... www.linkedin.com/pulse/quelqu...

linkedin.comQuelques réflexions sur l'"IApocalypse" à l'heure de l'entrée en bourse d'Anthropic« Combien vaut une entreprise qui pourrait éradiquer l’humanité ? », demandait cette semaine Nicolas Chapuis dans Le Monde. La question n’est pas seulement rhétorique.

Caroline Lequesne reposted

Caroline Lequesne reposted

"An accessible tool does not have to be an intrusive one," writes Ariana Aboulafia of CDT. She argues Meta used disability rights as a Trojan horse for facial recognition that threatens the privacy of people with and without disabilities.

techpolicy.pressFor Meta’s Smart Glasses, Access Is An AfterthoughtTech companies are using accessibility as a smoke screen to obfuscate the risks of AI wearables, writes Ariana Aboulafia.

Caroline Lequesne reposted

Caroline Lequesne reposted

Caroline Lequesne reposted

Incident reporting under the EU's AI Act will become binding in August, but the draft guidance still misses multi-agent and cascading AI failures, write Natàlia Fernández Ashman, Usman Anwar, and Dr. Marta Bieńkiewicz, urging the European Commission to close the accountability gap.

buff.lyEU Regulations Are Not Ready for Multi-Agent AI Incidents | TechPolicy.PressWith the AI Act's Article 73 guidelines set to become binding in August, the clock is ticking for the European Commission to address loopholes.

Caroline Lequesne reposted

Le ministre de l'intérieur a confirmé ce matin que les mégas fichiers du « Traitement d'antécédents judiciaires » (TAJ) et du « Fichier des personnes recherchées » (FPR) ont bien été compromis et que des données ont été volées. La Quadrature du Net s'alarme d'une catastrophe annoncée. ⬇️

Si la promesse est séduisante, on se souviendra tout de même de l'enquête MIT/Oxford il y a quelques mois qui nuançait ce genre de résultats "en condition". Cela ne nous épargne pas, en outre, des risques que la technologie fait peser sur les droits fondamentaux. www.journaldunet.com/intelligence...

journaldunet.comMAMMOth, ça ne trompe pas : enseignements d'une reconnaissance faciale sans biaisProjet européen MAMMOth : comment réduire les biais en reconnaissance faciale et améliorer la fiabilité, l'équité, la transparence et la conformité des systèmes d'IA.

Caroline Lequesne reposted

"The takedowns and restrictions began in October and targeted the Facebook, Instagram and WhatsApp accounts of more than 50 organisations worldwide ... in what appears to be a growing push by Meta to limit reproductive health and queer content across its platforms."

theguardian.comMeta shuts down global accounts linked to abortion advice and queer contentMore than 50 organisations report sites being restricted or removed, with abortion hotlines blocked and posts showing non-explicit nudity triggering warnings

Caroline Lequesne reposted

View full thread

Caroline Lequesne reposted

Caroline Lequesne reposted

"pushing risk down into the economy and onto adopters, reducing mechanisms that make risk visible and addressable, and stripping people of agency over their digital lives."

Caroline Lequesne@clequesne.bsky.social

« As leaked, the Omnibus would be the most significant and extraordinary retrenchment in digital rights in a generation » www.adalovelaceinstitute.org/news/ada-lov...

Caroline Lequesne reposted

This is very good and you should read it.

History Workshop@historyworkshop.org.uk

"Historical analysis is fundamentally different and more complex than producing a mass of visualizations and statistics that are the lifeblood of many A.I. programmes." Gordon McKelvie @gordonmckelvie.bsky.social on the problematic use of A.I. within historical research.

Caroline Lequesne reposted

The Trump Administration plans to use AI to make decisions about healthcare covered by Medicare. Miriam Osman, Emily Paul, and Emma Weil of Upturn discuss findings from their report that explores the use of similar automation in Medicaid, where it cut services and hid decisions behind algorithms.

buff.lyAutomation and the Fight for Care | TechPolicy.PressMiriam Osman, Emily Paul, and Emma Weil discuss how automation in homecare limits access to care and weakens healthcare rights.

Caroline Lequesne reposted

Criminal complaint against #facialrecognition company #Clearview #AI - noyb.eu/en/criminal-... "scraping billions of photos of Europeans and people around the world on the internet – and selling its facial recognition system to law enforcement and state actors" #gdpr

noyb.euCriminal complaint against facial recognition company Clearview AIClearview AI is known for scraping billions of photos of people around the world on the internet – and selling its facial recognition system to law enforcement and state actors

Caroline Lequesne reposted

Huge mess up by HMRC here but you also have to wonder which other organisations Home Office are sharing this error-ridden border data about individuals with?, how its being shared (copies? Held internally but queried?), whether limitations/errors are clear &c&c? www.theguardian.com/society/2025...

theguardian.comHMRC pauses child benefit crackdown after 23,500 families caught up in data errorParents say they were treated as fraudsters because Home Office travel records failed to show their return to UK

Caroline Lequesne reposted

« On est en train de discuter des livres que va lire Nicolas Sarkozy en prison ? Alors qu'il a été condamné pour "association de malfaiteurs". » @fabricearfi.bsky.social Journaliste d’investigation La suite : ⏰ 22h50 sur france·tv ➡️https://bit.ly/SarkozyEnPrisonHonneurDeshonneur 🎧en podcast

Caroline Lequesne reposted

#USA dans au moins 35 procès, les services gouvernementaux de l'administration US ont fourni aux Juges des informations fausses dont beaucoup de déclarations et de documents mensongers. Voilà ce qui nous attend si l'ED arrive au pouvoir (et Trump n'a même pas un an de mandat hein).

60 Minutes@60minutes.bsky.social

“We found over 35 cases in which the judges have specifically said what the government is providing…false information. It might be intentionally false information, including false sworn declarations time and again," says Ryan Goodman, law professor at New York University.