A new study reveals that leading AI laboratories have little to no publicly documented plans for identifying and containing AI models that behave in dangerous or unexpected ways. This gap in preparedness is drawing serious scrutiny as frontier models grow more capable and incidents of unexpected behavior become harder to dismiss.
Read the original →OpenAI is now publicly supporting stronger provisions in California's SB 53 AI safety legislation — a notable reversal from the company's earlier opposition to the bill. The shift signals a potentially significant realignment in how major AI developers are positioning themselves on state-level regulation.
Read more →The vast majority of published authors have had their work ingested into AI training datasets without consent or compensation, yet the legal picture remains murky. This deep-dive examines the competing fair use arguments, ongoing litigation, and what the outcome could mean for both the publishing industry and AI development.
Read more →A new AI model called Ox Alpha has emerged with no clear attribution, sparking intense speculation across AI communities about its origins and capabilities. The mystery surrounding its release has reignited broader conversations about anonymous or unaccountable AI development and the risks it poses.
Read more →British AI startup Inherent, founded by DeepMind alumni, has released an AI agent called Faraday that claims to outperform models from both Anthropic and OpenAI on the task of replicating published scientific papers. The ability to reliably reproduce research findings at scale could have significant implications for accelerating scientific discovery and validating existing work.
Read more →Flock Safety, a company that sells AI-powered license plate readers and surveillance networks to law enforcement, is facing mounting public criticism over potential misuse of its technology. The CEO's call for "compromise" comes as civil liberties advocates intensify pressure over the lack of oversight governing how the data is collected and shared.
Read more →Harvard Business School's new HBS Foundry program uses AI-generated avatars of its faculty to deliver real-time feedback during practice pitch sessions and simulated board meetings. The $699 price point and AI-augmented format represent a notable experiment in scaling elite business education beyond the traditional classroom.
Read more →Summaries are AI-generated. Click through to read the original reporting.
Summaries are AI-generated. Click through to read the original reporting.