The U.S. gated its newest AI models. A Chinese lab gave one away.

Issue

AI for FIs, 2026-06-30

Author

Brent Dixon

Topics

Open Source, Model Risk, Credit Unions, Infrastructure, Risk Management, Governance, Lending, Agentic Ai, Anthropic/Claude, Openai, Ai Safety

Plus: agents moving from scoring the loan to making the call, and a malicious AI tool that cleared review and reached 26,000 users.

• Read the full issue (canonical)
• Subscribe to AI for FIs