After Andrew White was granted access to GPT-4, the new artificial intelligence system that powers the popular ChatGPT chatbot, he used it to suggest an entirely new nerve agent.
The chemical engineering professor at the University of Rochester was among the 50 academics and experts hired to test the system last year by OpenAI, the Microsoft-backed company behind GPT-4. Over six months, this “red team” would “qualitatively probe [and] adversarially test” the new model, attempting to break it.
White told the Financial Times he had used GPT-4 to suggest a compound that could act as a chemical weapon and used “plug-ins” that fed the model with new sources of information, such as scientific papers and a directory of chemical manufacturers. The chatbot then even found a place to make it.