Find a briefing by day

OpenAI says AI is reshaping cybersecurity for attackers and defenders. It also says it is strengthening its own defenses and sharing actions teams can take now.

OpenAI says its custom “Jalapeño” chip runs AI responses faster and with less power use, aiming for quicker replies and more work done at once.

Google DeepMind says it is piloting “double-blind” evaluations for AI, aiming to reduce bias in how AI systems are judged and compared.

OpenAI and Thailand’s MHESI are starting an eight-week program to help 10 startups in health, wellness, and education move from demos to trusted products.

OpenAI says it will expand its presence in Brazil and deepen engagement with local builders, businesses, and communities to support AI use.

OpenAI says it investigated a security incident involving Hugging Face and is taking steps to strengthen AI security, monitoring, and safety behavior.

OpenAI says it banned Russia-origin accounts using AI to promote a fake Israel-based think tank and a “sovereignty” index praising Russia and criticizing the West.

OpenAI says it is strengthening monitoring, safety work, and security for its most advanced AI systems, and using new safeguards to guide release timing.

Hugging Face published a post on how to measure “benchmark optimization” in speech recognition, focused on detecting when systems improve on tests without real progress.

Researchers propose a video workflow where teachers and automated checks both reject weak outputs, aiming to keep videos instructionally sound, not just polished.

OpenAI says Stampli used Codex and ChatGPT Work to compress weeks of launch production into days amid a fixed deadline and limited design resources.

OpenAI says a new system will look for risky patterns across related API interactions while keeping eligible customers' prompts unavailable to its staff.