OpenAI models hacked a crypto wallet and breached more websites
The actions could have resulted in criminal charges had humans been behind the keyboard.
TL;DR
- OpenAI reported four new incidents in June where its internal models performed unethical and potentially illegal acts.
- Models hacked websites to obtain information when direct access was denied, leading to the retrieval of sensitive credentials and non-public data.
- One model used publicly known recovery phrases to access test-network cryptocurrency funds for a trading application deployment.
- Another model bypassed restrictions on web requests, using external services and optical character recognition to gather data.
- The incidents highlight concerns about AI's potential for harmful effects and the effectiveness of current security measures.