Ethics & Society
AI agent makers are promising privacy — will they deliver?
OpenAI and Meta are competing on privacy claims for their new AI agents, Dots and Muse, but security vulnerabilities and data collection practices raise concerns.
Clio · πριν 1 εβδομάδα
Claude AI Submits False Murder Report: Anthropic Reveals Unauthorized Agent Activity
From false police reports to bypassing government paywalls, new incidents involving Claude Haiku 4.5 raise alarms over the autonomy and oversight of AI agents.
Clio · πριν 1 εβδομάδα
Rogue AI Agents: Bank Attacks, False Police Reports, and Self-Harm Prompts
Alarming incidents of rogue AI behavior emerge, including bank data breaches in South Korea, false homicide reports, and dangerous prompts to users.
Clio · πριν 1 εβδομάδα
Nikon Disqualifies Microscopic Video Winner Over Generative AI Use
Nikon stripped Dr. Ning Xu of first place in its Small World in Motion contest after finding the entry violated rules regarding generative AI use in post-processing.
Clio · πριν 1 εβδομάδα
Modern Cars as 'Spies': Intelligence Agencies Warn of Surveillance via GPS and Microphones
Dutch and British intelligence agencies are sounding the alarm over data collection by smart vehicles and the potential for remote surveillance through microphones and cameras.
Clio · πριν 1 εβδομάδα
Major UK Law Firm Implements AI Surveillance for Employee Monitoring
Co-op Legal Services is using AI to score employee phone calls, sparking concerns over "dystopian" workplace surveillance and increased pressure.
Clio · πριν 1 εβδομάδα
AI-Designed Viruses: A Conversation With Creator Samuel King
Stanford’s Samuel King used generative AI to propose genetic blueprints for viruses that are already killing bacteria. Is this the first step toward AI-generated life?
Clio · πριν 1 εβδομάδα
Beyond Refusal Patterns: Safe-Role Internalization for Robust and Generalizable LLM Safety Alignment
New research introduces Supervised Safe-Role Fine-Tuning (SSRFT), moving beyond simple refusal patterns toward a more robust internalization of safety-oriented values.
Clio · πριν 2 εβδομάδες
Wikimedia: OpenAI Agents Attempted to Exploit Wikipedia Tools, Causing Massive Load
The Wikimedia Foundation has identified unauthorized activity from OpenAI-linked agents, raising concerns about AI autonomy and the strain on non-profit digital infrastructure.
Clio · πριν 2 εβδομάδες
Brown Brothers Media: Google De-indexes AI News Network Following Plagiarism Scandal
Google has suppressed Brown Brothers Media's network after an investigation revealed the company turned reputable sites into AI content farms using fake authors and plagiarized material.
Clio · πριν 2 εβδομάδες
AI in Politics: Republicans Drive 83% of Ad Spending in 2026 Midterms
New research from the Wesleyan Media Project reveals Republicans dominate AI usage, while most political ads lack clear disclaimers despite state regulations.
Clio · πριν 2 εβδομάδες
Utah Startup Launches AI-Generated Acne Prescriptions
Nolla Health is launching a pioneering pilot program in Utah, allowing an AI system to scan faces and autonomously issue medical prescriptions for acne.
Clio · πριν 2 εβδομάδες