OpenAI has released a new benchmark for testing AI systems in healthcare. Called HealthBench, it's designed to evaluate how well language models handle realistic medical conversations. According to ...

New research shows that AI agents with internet access are vulnerable to simple manipulation tactics. Attackers can deceive these systems into revealing private information, downloading malicious ...

The Trump administration is preparing a regulation that would force AI companies with federal contracts to maintain strict political neutrality. According to the Wall Street Journal, the move targets ...

According to OpenAI's "State of Enterprise AI 2025" report, ChatGPT Enterprise users save an average of 40 to 60 minutes per active workday. Workers in data science, engineering, and communications ...

Anthropic is treating its new Claude Opus 4 language model as safety-critical after tests revealed some troubling behavior, including escape attempts, blackmail, and autonomous whistleblowing. After ...

Salesforce's new CRMArena-Pro benchmark reveals major challenges for AI agents in business contexts. Even top models like Gemini 2.5 Pro manage just a 58 percent success rate on single turns. When the ...

Sam Altman, OpenAI's CEO, shared his ideas about the future of AI interaction at the company's DevDays event. In a talk with Chief Product Officer Kevin Weil, Altman described an AI system that could ...

xAI has officially announced Aurora on its blog, confirming it as an entirely new model built from the ground up. This suggests the company may be moving away from its previous partnership with Black ...

A flawed analysis by Goldman Sachs on ChatGPT's traffic decline caused market jitters. But the data was misread. Demand for OpenAI's services continues to grow. A report by Goldman Sachs analyst Peter ...

Instead of playing a proper game of chess against Stockfish, a dedicated chess engine, o1-preview figured out how to hack its test environment to force a win. According to Palisade Research, an AI ...

While not explicitly named, China (including Hong Kong), Russia, North Korea, and Iran are likely to be the main-affected countries. These nations are notably absent from OpenAI's list of supported ...

The recently introduced Gen-4.5 now features native audio generation and audio editing, as well as multi-shot editing, a feature that lets users apply changes to a single scene and have them ripple ...