OpenAI says its automated research intern can complete well-defined AI research tasks that would take a skilled researcher ...

OpenAI found GPT-6 Astra harder to monitor through its chain of thought than GPT-5.6 Sol, especially when told to evade ...

OpenAI classified Astra as its first model to reach a Critical cybersecurity threshold after finding and exploiting unknown ...

Meta tested smaller AI-supported teams as software changes rose 220% year-on-year, compared with a 36% increase in changes ...

AI-enabled fraud is outpacing traditional controls, increasing the need for integrated data and faster fraud detection.

The two announcements, made at Equinix Horizon, include Equinix Inference Exchange, developed with NVIDIA and Together AI, ...

Baidu’s Apollo Go robotaxi expansion into Southeast Asia expected to target Singapore and Malaysia. Chinese robotaxi firms lead globally with 11+ million rides, US rivals focus on domestic markets. As ...

TikTok has agreed to an Oracle-led plan to move its US business into a new joint venture. ByteDance’s stake would shrink, but questions over control and compliance remain. For years, the future of ...

The spotlight this week has been on AI chips — and the growing push to make them more efficient. “That helps to max out TSMC technology’s capability, and we find this is very useful,” said Jim Chang, ...

Google Cloud unveils tools to secure AI agents at 2025 Security Summit. Updates help keep AI projects compliant. Google Cloud used its Security Summit 2025 to highlight a wave of new tools aimed at ...

Anthropic Claude is now available in Microsoft Foundry with Azure controls. The rollout includes NVIDIA GB300 infrastructure and Foundry Agent Service support. Anthropic’s Claude models are now ...