Weekly Update: Week of July 21, 2025

Comprehensive AI policy shifts. This week culminated in major policy announcements. On July 23, the White House unveiled “America’s AI Action Plan,” outlining the U.S. administration’s strategy for AI. The plan strikingly prioritized deregulation and innovation, directing agencies to eliminate or revise rules seen as impediments to AI growth. It even instructed NIST to remove references to “misinformation, diversity, equity, and inclusion” from its AI Risk Management Framework to avoid “social engineering agendas”. Simultaneously, President Trump issued executive orders barring the federal government from using any AI deemed “politically biased” – part of an effort against so-called “woke AI”. These moves signal a clear federal pivot to an industry-friendly, laissez-faire approach, raising concerns among safety advocates. On the innovation front, OpenAI announced plans to launch GPT-5 in August, an experimental next-generation model expected to push the envelope of AI capabilities. The company hinted at new research techniques in GPT-5, while noting it won’t include a contentious math-solving module that had recently propelled its AI to Olympiad gold. This disclosure fed into ongoing regulatory discussions: agencies like the FTC and European regulators are anxious to scrutinize new ultra-powerful models for safety issues before deployment. The final week of our two-month review thus encapsified the dynamic tension in AI governance – bold pro-innovation policies on one hand, and parallel efforts to plan oversight of the next AI breakthroughs on the other.

(In sum, the last two months saw unprecedented activity in AI governance. Multiple U.S. states advanced AI laws, the EU finalized its landmark AI Act, and global bodies from the U.N. to G7 ramped up involvement. While approaches differ, the common goal is to reduce AI’s risks – whether through transparency, accountability, or outright restrictions – while still encouraging beneficial innovation.)

Sources: Highlights compiled from Securiti’s Global AI Regulation Roundup (May–July 2025), Reuters and Bloomberg news, and official government releases (White House, EU Commission).