AI Agents in Fintech: Are They Out of Control?

Are AI agents out of control, or are most companies just governing them with suggestions? A solo episode on agentic AI in banking and fintech: who is in control, and who is accountable when an agent gets it wrong. In July, hundreds of AI agents inside an OpenAI test broke into Hugging Face. Meanwhile TD Bank cut a mortgage step from fifteen hours to under three minutes with an agent. Same technology, opposite outcomes. This episode looks at why, what my podcast guests disagree on, and why US bank model-risk guidance still excludes agentic AI by name. Chapters 0:00 Hundreds of AI agents broke into Hugging Face 0:50 Risk or productivity tool? 2:15 What my guests said 3:42 My take: prompts are suggestions, not controls 5:36 The evidence on both sides 7:52 What it means for banks and fintechs Episodes mentioned Parul Kaul-Green on AI in insurance claims:    • The END of Slow Insurance? AI Underwriting...   Krista Griggs on trust and technology in digital banking:    • Digital Banking Future: Trust vs Tech | Kr...   Christie Kristensen on open banking APIs:    • Open Banking APIs Transform Fintech | Chri...   Sources Hugging Face technical timeline: https://huggingface.co/blog/agent-int... OCC Bulletin 2026-13, model risk management: https://www.occ.gov/news-issuances/bu... METR study on AI tools and developer speed: https://metr.org/blog/2025-07-10-earl... 🎙️ Subscribe for weekly episodes on fintech, AI, and innovation 🌐 Website: https://curiositycodepodcast.com/ 📱 Follow on LinkedIn:   / alexkhomyakov   🎧 Listen on Spotify: https://open.spotify.com/show/4ePKf2q... 🍎 Apple Podcasts: https://podcasts.apple.com/podcast/id... Agentic AI is moving into lending, fraud and payments faster than the rules that govern it. For anyone running technology at a bank or fintech, the question is where your agents' limits actually live, and who owns them. #AIAgents #AgenticAI #Fintech #AIGovernance #Banking