Evolving AI may arrive before AGI and create hard-to-control risks
(Left) Humanoid robot Alter3 equipped with a Society of Mind architecture based on LLMs for communication and operation…
Is artificial general intelligence already here? A new case that today’s LLMs meet key tests
Credit: Pixabay/CC0 Public Domain Will artificial intelligence ever be able to reason, learn, and solve problems at levels…
Agentic AI threatens research funding system
Credit: David Parkins, from Nature (2026). DOI: 10.1038/d41586-026-01297-y In a new analysis, two UCL researchers argue that the…
EU tells Google to open Android to AI rivals
The European Commission wants consumers to be able to try more than just Gemini — Google’s AI app…
Google commits $10 billion, could add $30 billion more to Anthropic
Anthropic CEO Dario Amodei has visited the White House as the startup stands its ground regarding safe use…
OpenAI launches GPT-5.5 as rivals race to build more autonomous AI assistants
OpenAI president Greg Brockman says the new GPT-5.5 model can tend to more computer work without human supervision.…
Sharper bias tests could help stop ChatGPT from amplifying hidden stereotypes
Credit: Matheus Bertelli from Pexels Language models like ChatGPT are not neutral. Without our realizing it, they can…
Restricted Mythos AI model reaches unauthorized users, triggering Anthropic probe
Anthropic has delayed a general release of its latest model Mythos, which it says can spot undiscovered security holes…
Amazon invests another $5 bn in Anthropic
As Amazon invests billions of dollars in Anthropic, the AI startup commits to spending heavily on the tech…
Anthropic says will put AI risks ‘on the table’ with Mythos model
Claude Mythos’ release has been restricted to a small number of partners. American AI developer Anthropic plans to…
Unpredictable AGI may resist full control, making diverse AI safer
Credit: King’s College London Public concern about AI safety has grown significantly in recent years. As AI systems…
This AI mines the numbers buried in scientific papers and turns them into usable data fast
The Quinex framework, developed by researchers at Jülich, is based on language models and automatically identifies numerical values…