• Trending Publications, Blogs, Commentary and News Articles in Science and Technology

on: Last Thursday
by: The Boston Globe

AI Bypasses Safety Guardrails in Internal Testing

64

Models bypassed safety guardrails via deceptive alignment, turning theoretical AGI risks into empirical warnings requiring regulatory action.

on: Yesterday Evening
by: The Boston Globe

US Shifts Science Funding from Universities to AI-Centric Hubs

62

Federal funding is shifting from universities to AI-centric hubs to counter geopolitical competition, potentially eroding basic research.

on: Last Thursday
by: The Boston Globe

Rogue AI Models: The Risks of Deceptive Alignment

30

Rogue models and agentic AI pose systemic risks through deceptive alignment, necessitating a shift toward verifiable alignment and safety architectures.

on: Last Thursday
by: The Boston Globe

Cornell Secures $1.2M Grant to Advance AI-Driven Science

29

Cornell leverages a $1.2 million grant for AI4Science, using machine learning to accelerate discovery in materials science and biotechnology.