L'IA transforme les publicités Facebook en profils personnels

10 mai 2026

Rapporté par l'IA

L'intelligence artificielle peut désormais extraire des profils personnels complets à partir des publicités Facebook, plus rapidement et à moindre coût que prévu. Des publicités anodines semblent contenir suffisamment de données pour cartographier les identités de manière détaillée.

Cette capacité transforme des publicités ordinaires en sources d'informations détaillées sur les utilisateurs. Ce développement soulève des questions sur la manière dont les données issues des plateformes sociales peuvent être traitées de nouvelles manières.

Articles connexes

Study shows AI can deanonymize online users from posts

A new research paper demonstrates that large language models can identify real identities behind anonymous online usernames with high accuracy. The method, costing as little as $4 per person, analyzes posts for clues and cross-references them across the internet. Researchers from ETH Zurich, Anthropic, and MATS warn of reduced online privacy.

Hackers use AI to exploit security flaws faster, IBM finds

26 février 2026 Rapporté par l'IA

Hackers are increasingly leveraging artificial intelligence to identify and exploit security vulnerabilities at an accelerated pace. According to a report from IBM, the integration of AI into cyber attacks is speeding up the process significantly. This development highlights evolving threats in cybersecurity.

Technologie

Google adds reddit advice and more sources to ai search

Politique

Artificial intelligence enhances 2026 municipal elections for better and worse

Technologie

Snapchat rolls out sponsored AI agents in chat

Fake Chrome AI extensions targeted over 300,000 users

Criminals have distributed fake AI extensions in the Google Chrome Web Store to target more than 300,000 users. These tools aim to steal emails, personal data, and other information. The issue highlights ongoing efforts to push surveillance software through legitimate channels.

Meta develops AI version of Mark Zuckerberg for employees

13 avril 2026 Rapporté par l'IA

Meta is creating an artificial intelligence version of its chief executive, Mark Zuckerberg, to interact with employees. The project involves photorealistic 3D characters trained on Zuckerberg's mannerisms, tone, and statements. Zuckerberg is personally training and testing the animated AI as part of the company's AI push.

New cybercrime platform 1Campaign aids malicious Google ads

A new cybercrime platform known as 1Campaign allows hackers to run malicious Google Ads while evading the company's screening process. This development is raising concerns in the cybersecurity community. The platform's emergence highlights ongoing challenges in online advertising security.

dimanche 03 mai 2026 16h34