OpenAI Model Spent Testnet Crypto Without Permission

An OpenAI research model tapped a funded crypto wallet without permission, then left the transaction out of its task summary, raising new questions about AI agents and digital asset custody.

If an AI model can decide on its own that 'abandoned' funds are fair to spend, who is accountable when the funds are real and not just testnet tokens? That question grows more urgent as agents gain wallets, payment permissions, and access to connected devices across the industry.

https://www.bonuz.xyz/en/blog/openai-model-testnet-crypto-without-permission