OpenAI Is Experimenting With AI in Bioscience Research at the Lab That Built the First Nuclear Weapon

In a July 9 statement, OpenAI said it was working with the laboratory’s bioscience division to evaluate how AI models such as GPT-4o can assist scientists

OpenAI is partnering with Los Alamos National Laboratory (LANL) — the lab responsible for building the first nuclear weapon in the world — to experiment with the use of AI in bioscience research.

In a July 9 statement, OpenAI said it was working with the laboratory’s bioscience division to evaluate how AI models such as GPT-4o can assist scientists with performing tasks in a physical lab.

“This includes biological safety evaluations for GPT-4o and its currently unreleased real-time voice systems to understand how they could be used to support research in bioscience,” wrote OpenAI.

LANL was founded in 1943 with a single goal in mind, to design and build the world’s first atomic bomb. For nearly 50 years, the lab focused primarily on high-level military research but following the end of the Cold War in 1991 it shifted its focus to civilian research and development.

Today, the bioscience division of LANL works on a wide range of research areas, including vaccine development, sustainability biotech, the impact of climate change on disease emergence, and various forms of biothreat detection.

GPT4o is the latest “multimodal” large language model from OpenAI, allowing users to converse with the chatbot in real-time with voice, imagery, and text.

The partnership’s mandate outlines tasks such as cell transformation, which involves introducing foreign genetic material into a host organism, cell culture, and cell separation.

OpenAI said the upcoming partnership will build on its previous bioscience work in two primary dimensions, including the incorporation of “wet lab techniques,” which involves training models to assist in conducting more complex tasks like mass spectrometry.

Additionally, the AI firm said its work would focus on new “modalities” of AI use in the laboratory setting, with GPT-4o’s voice and visual inputs opening up new channels to “potentially expedite learning” among scientists.

The artificial intelligence firm said the work with the laboratory would contribute to its ongoing work on biothreat risks.

The new partnership comes as large tech companies — including Microsoft and Apple — begin to distance themselves from the AI startup amid growing regulatory concerns.

On July 10, Microsoft penned a letter to OpenAI announcing its withdrawal from the board, just one year after the Windows software maker made a $13 billion investment in the firm.

In June, European Union regulators announced that OpenAI could face an EU antitrust investigation over its partnership with Microsoft.