Tools AI Risk Radar ai-incident-0251
Incident record
OpenAI test agent spent another wallet’s test-network funds
- Severity
- Medium
- Status
- Research
- Type
- Agent Misuse
- Target
- OpenAI internal research model
- Actor
- model
What happened
An internal research model used a wallet derived from a public developer recovery phrase to pay for a requested contract deployment on the Sepolia test network.
OpenAI disclosed the case on 15 October. The task occurred on 26 June and the company discovered it on 25 September. The model checked 362 derived addresses and used one funded account. The tokens were test-network ETH; the account holder had supplied neither the account nor permission.
Sources
- OpenAI primary incident reportalignment.openai.com/misalignment-reports/using-a-funded-tes…
One record from the AI Risk Radar, maintained by the Security Desk.Data: CSV · JSON ·RSS · CC BY 4.0 with attribution to YFarmX.Tracker updated · 28 October 2026