Escalating Global Scrutiny and Geopolitical Race in AI Safety and Governance

Key Intelligence

• ■Calls for AI safety and robust oversight are escalating globally, with proposals ranging from a 'hardwired pause' on frontier AI training to California's new executive order proposing a 'kill switch' for high-risk applications.
• ■The US-China AI rivalry continues to intensify, with China's energy advantages potentially fueling its pursuit of dominance, while concerns persist over its minimal publication of AI model safety tests (only 3.6%).
• ■Ethical dilemmas abound as AI is increasingly integrated, from questions about human perception of AI and its efficacy in mental health support to challenges in maintaining information integrity and addressing copyright and workforce impacts (e.g., publishers' staff revolting).
• ■Regulatory frameworks are evolving, with legal decisions clarifying fair use for AI training data and ongoing updates to data privacy laws like GDPR, indicating a growing legal landscape for AI governance.