OpenAI Partners with Los Alamos to Study AI in Lab Settings

OpenAI and Los Alamos National Laboratory are collaborating to explore the benefits and risks of using generative AI in an active laboratory setting, as detailed to Axios. OpenAI CEO Sam Altman has emphasized generative AI’s potential to advance human understanding in science, despite concerns about its misuse, such as the creation of bioweapons.

Initial Experiment: AI in Biomedical Tasks

The initial experiment involves using AI to assist individuals without expertise in molecular biology to perform basic biomedical tasks. Specifically, AI will help genetically engineered E. coli bacteria produce insulin. OpenAI believes this is the first experiment of its kind and aims to identify which research areas genAI systems can best support.

Real-World Application and Evaluation

Tejal Patwardhan, a technical lead on OpenAI's preparedness team, highlighted the importance of testing advanced models in real-world settings. “We want these advanced models to contribute to scientific progress — and evaluating them in a setting where scientists actually work is one of the first steps toward realizing their potential,” Patwardhan told Axios.

Balancing Benefits and Risks

Nick Generous, a deputy group leader at Los Alamos, noted the importance of understanding AI's role in the lab both as a tool and a potential threat. “Many great things can happen, but we want to make sure that if any risks do get introduced that we can at least identify and mitigate them before they become a problem,” Generous said.

Focus on Practical Application

Previous research by OpenAI’s team found that GPT-4 provided a "mild uplift" in delivering information that could potentially lead to biological threats. However, those experiments were limited to written tasks. The collaboration with Los Alamos will focus on integrating text, vision, and voice data to explore AI’s practical applications in the lab.