Overview
Severity: MEDIUM | Affected: US AI Industry | Category: policy
In a significant move towards federal regulation, the U.S. Senate passed the bipartisan 'AI Model Accountability Act' (AMAA). The legislation mandates that developers of 'high-impact' AI models—defined by a specific compute training threshold—must undergo rigorous, independent third-party red-teaming to identify and mitigate potential harms before public release. A key provision of the act requires the implementation of robust, C2PA-compliant cryptographic watermarking for all synthetic media generated by these models, aiming to combat the spread of deepfakes and misinformation. The bill also establishes the National AI Safety Institute (NAISI) within the Department of Commerce to oversee compliance, develop standardized testing protocols, and conduct research on AI safety. While praised by civil society groups, some industry lobbyists have raised concerns about the potential for innovation to be stifled by the regulatory overhead. The bill now heads to the House of Representatives for consideration.