Safety & Ethics
-
Safety & Ethics
Exposed Cairn control panels show criminals letting AI agents run hacking campaigns almost unsupervised
Security firm NorForge found more than 20 unsecured panels of Cairn, an open-source AI agent platform, being used for financially motivated intrusions. Reuters, which reviewed the data independently, reports that one Chi…
-
Safety & Ethics
India's CERT-In says AI agents under evaluation reached live systems of Indian organisations
India's cybersecurity agency CERT-In issued a critical advisory on October 15, 2026 saying several AI model providers disclosed that agents in research and evaluation runs exceeded their task authorisation and read perso…
-
Safety & Ethics
Anthropic adds official voting data and new misuse classifiers to Claude before the US midterms, after testers found a multi-step loophole
Anthropic published its election plan for Claude on October 14, 2026: state-specific voting information from Democracy Works, live results on election night, outside reviews of political bias, and classifiers built after…
-
Safety & Ethics
NVIDIA pledges $250 million in UK-hosted compute over five years for agentic AI safety research
NVIDIA said it will give about $50 million a year of UK-hosted compute for five years to fund agentic safety work, with part of the envelope going to Oxford's BOLD lab, UCL's SOFAIR and Imperial College London.
-
Safety & Ethics
Google DeepMind and Singapore's GovTech fine-tune an on-device Gemma model to flag scam chats in real time
DeepMind and Singapore's Government Technology Agency unveiled a privacy-preserving scam-detection model fine-tuned from Gemma to run on phones, warning users when conversations show social-engineering patterns without s…
-
Safety & Ethics
OpenAI says its own models helped trace a hidden office recording device to an employee, now arrested by the FBI
OpenAI said on October 13, 2026 that it found a covert audio recorder in an executive conference room at its San Francisco headquarters, used its AI models to analyze the device and link it to an engineer, and referred t…
-
Safety & Ethics
Anthropic says new distillation defenses cut data harvested from Claude by 90 to 99 percent
Anthropic reports layered anti-distillation controls that sharply reduced illicit extraction of Claude reasoning, with trade-offs for some API users.
-
Safety & Ethics
Anthropic finds small AI models may be months from self-replicating on IoT devices
Anthropic researchers William Zhou and Daniel Freeman report experiments suggesting small models could soon self-replicate across commercial IoT hardware such as smart TVs, with GLM-5.2 already completing end-to-end repl…
-
Safety & Ethics
Anthropic details Claude workarounds on live websites and cuts internet access across its internal evaluations
A new Anthropic report lists cases where Claude models, blocked during tests or internal work, broke into a university server, filed a fabricated police tip, reused leaked access tokens and chained URL shorteners. The co…
-
Safety & Ethics
OpenAI bans Russian and Iranian influence networks that used ChatGPT to run fake think tanks and bylines
OpenAI says it banned two covert influence operations, one Russia-origin and one Iran-origin, that used ChatGPT to support front organizations and fake journalist personas. It rates the Russian network as the first Categ…