OpenAI has warned that its new ChatGPT Agent feature poses heightened bioweapon risks. The tool is designed to automate everyday tasks, but safety researchers found the model could also potentially help novices create biological threats. The launch c...