Countries sidestep AI safety obligations at global AI summit.

This weekly news briefing covers international AI policy and several security stories: a non-binding AI declaration at India’s AI Impact Summit reflecting a shift toward AI adoption over safety; reporting that Chinese firm DeepSeek may have trained a frontier model on U.S.-restricted NVIDIA Blackwell chips, raising export-control and national-security concerns; and separate headlines including an indictment of three former engineers accused of stealing Google trade secrets, U.S. defense cybersecurity rules impacting small suppliers, and a West Virginia lawsuit against Apple over alleged CSAM distribution.