Rocket Lab’s media mentions surged 40-fold, marking a major increase in its global visibility and coverage, according to GDELT data.

Anthropic introduces an invisible watermark for future Claude models, using a secret key to signal AI involvement without revealing authorship or adding visible marks.

OpenAI has temporarily slowed Astra model development and paused reinforcement learning after preliminary evidence suggested cybersecurity risks, with ongoing safety tests.

Analysis of China’s focus on end-to-end AI solutions for exports, highlighting potential impacts on global markets and industry strategies.

OpenAI launches a $5 million program to support democratic oversight bodies in reviewing government use of AI in national security, emphasizing transparency and accountability.

OpenAI reports Asana completed five years of engineering work in two weeks with Codex, raising questions about the scope and verification of the claim.