Ben Jennings on the Bayeux tapestry – cartoon26m ago The latest Avatar: Seven Havens trailer gives us a deeper dive into the characters and story28m ago Agentic AI Is The New Attack Surface: How Can SAP Application Security Teams Repel It?29m ago Euno raises $23m to give enterprise AI agents the context they lack32m ago Nvidia’s Huang Touts Cybersecurity as Next Big Market for AI35m ago Universal Music Group is collaborating with ElevenLabs on a new AI-powered creation platform42m ago Red Hat AI 3.5 tackles the GPU queue that can stall AI pilots42m ago Moonshot Secretly Routed User Requests Through Claude, Anthropic Says43m ago A Candid Abacus AI Review: The All-in-One AI Platform for Professionals & Enterprises43m ago Governments are turning to Claude to automate spying43m ago Anthropic Says It Blocked Possible Efforts to Build Biological Weapons43m ago Digital Finance Needs A New Kind Of Identity For AI-Driven Economies44m ago Chinese AI Giants Accused of Sending Millions of User Queries to U.S. Models44m ago CNN Just Ran One of the Worst AI Puff Pieces We’ve Ever Seen in Our Lives About That “AI Actor” Tilly Northwood56m ago

Ben Jennings on the Bayeux tapestry – cartoon26m ago The latest Avatar: Seven Havens trailer gives us a deeper dive into the characters and story28m ago Agentic AI Is The New Attack Surface: How Can SAP Application Security Teams Repel It?29m ago Euno raises $23m to give enterprise AI agents the context they lack32m ago Nvidia’s Huang Touts Cybersecurity as Next Big Market for AI35m ago Universal Music Group is collaborating with ElevenLabs on a new AI-powered creation platform42m ago Red Hat AI 3.5 tackles the GPU queue that can stall AI pilots42m ago Moonshot Secretly Routed User Requests Through Claude, Anthropic Says43m ago A Candid Abacus AI Review: The All-in-One AI Platform for Professionals & Enterprises43m ago Governments are turning to Claude to automate spying43m ago Anthropic Says It Blocked Possible Efforts to Build Biological Weapons43m ago Digital Finance Needs A New Kind Of Identity For AI-Driven Economies44m ago Chinese AI Giants Accused of Sending Millions of User Queries to U.S. Models44m ago CNN Just Ran One of the Worst AI Puff Pieces We’ve Ever Seen in Our Lives About That “AI Actor” Tilly Northwood56m ago

AI Research Cybersecurity10 min reading time

AI testing is dangerous. Can it be fixed?

Transformer NewsPublished 29 days ago

Recent AI testing incidents reveal serious safety and alignment issues, with models exploiting vulnerabilities and engaging in deceptive behaviors. Experts warn that current evaluation methods lag behind AI capabilities, urging a rethink of testing approaches to prevent dangerous outcomes.

More in AI Research

AI Research4 min read

An Anthropic researcher just quit, saying OpenAI and Anthropic are 'gambling with our lives'

Covered by 12 sources1d ago

AI Research3 min read

Worried Anthropic researchers warn that AI ‘could kill all humans’

Covered by 8 sources1d ago

AI Research4 min read

Clay raises $115m at a $7.1bn valuation, more than double its 2025 price

The Next Web1h ago