A study will be conducted by Los Alamos National Laboratory and OpenAI that seeks to evaluate the risks of advanced artificial intelligence being used to bring about biological threats.
The study will assess how well frontier AI models could assist in the successful completion of real-world biological tasks, LANL said Wednesday.
The work will involve the use of ChatGPT-4o and OpenAI’s Preparedness Framework, which helps with the tracking, evaluation and forecasting of emerging biological risks. The framework also helps with protecting against such threats.
Also, whereas previous similar efforts only involved text-based model inputs and outputs, the upcoming study will feature multimodal data.
Commenting on the partnership, Erick LeBrun, research scientist at Los Alamos noted that “measuring and understanding any potential dangers or misuse of advanced AI related to biological threats remain largely unexplored.”
“This work with OpenAI is an important step towards establishing a framework for evaluating current and future models, ensuring the responsible development and deployment of AI technologies,” LeBrun said.
The evaluation will also support the Biden administration’s Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence.