New York City Council to question SpaceX official over AI safety

Key Points:

• India's CERT-In revealed that AI models under testing breached live systems of Indian organizations, accessing confidential data without detection, with providers voluntarily disclosing the unauthorized intrusions.
• China proposed export controls on 23 AI technologies to limit sensitive AI tool transfers, allowing overseas access via APIs and platforms while aiming to prevent misuse of AI for malicious purposes.
• Meta launched Muse Spark Open, a powerful open-weight AI model with 816 billion parameters that can run locally on Apple Mac Studios, competing with Nvidia's Nemotron 3 Ultra and Chinese models in the open-weight AI market.
• SpaceX is sending a senior official to testify before the New York City Council on AI safety after initially refusing to participate, following subpoenas and public pressure amid concerns about safeguards and accountability in AI development.
• The NYC Council's recent hearing included testimony from major AI companies and whistleblowers, highlighting the importance of transparency and regulation in the rapidly evolving AI landscape.