Forging private and public sector collaboration, OpenAI and Los Alamos National Laboratory (LANL) announced a partnership to evaluate the applications and risks of artificial intelligence (AI) in bioscience.

“As a private company dedicated to serving the public interest, we’re thrilled to announce a first-of-its-kind partnership with Los Alamos National Laboratory to study bioscience capabilities. This partnership marks a natural progression in our mission, advancing scientific research while also understanding and mitigating risks,” said OpenAI’s Chief Technology Officer Mira Murati.

The partnership seeks to establish standards for the use and safety of multimodal AI models in the sciences, giving way for innovations and advancements in bioscience, healthcare, and other areas.  

LANL is the next among a list of public entities, such as Moderna and Color Health, that have teamed up with private-owned company OpenAI to leverage its technology in building AI-supported assistants, following the White House Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence.

Testing the Capabilities and Risks of AI

OpenAI claimed that its collaboration with LANL will be the “first experiment to test multimodal frontier models in a lab setting.” The company’s frontier AI models, GPT-4o and its unreleased real-time voice system, will assist both expert and amateur scientists in performing standard experimental tasks that will cover transformation, which involves introducing foreign genetic material into a host, cell culture, and cell separation, among other things.

These multimodal technologies will also be investigated to assess and understand the risks of their misuse. Researchers at LANL have created an AI Risks Technical Assessment Group, which will focus on how OpenAI’s generative AI can become a weapon to create biological threats.  

“The potential upside to growing AI capabilities is endless. However, measuring and understanding any potential dangers or misuse of advanced AI related to biological threats remain largely unexplored,” stated LANL scientist Erick LeBrun. “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.”

Exploring New Dimensions

Besides the promising impacts of artificial technology on the future of bioscience, OpenAI upholds that the upcoming partnership will also open opportunities for the company’s previous work to reach new dimensions and provide better service to customers.

While ChatGPT and GPT-4 can provide texts detailing tasks and responses to questions involving scientific benchwork, they do not fully capture the skills needed to enable users to perform biological experiments. OpenAI believes that incorporating wet lab techniques in training its models will enable them to assist better in accomplishing more complicated tasks in the laboratory, such as mass spectrometry.

Moreover, by using GPT-40’s capability to function in multiple modalities through both text and audio inputs, the company is hopeful that learning can become faster and more effective. “For example, a user less familiar with all the components of a wet lab setup can simply show their setup to GPT-4o and prompt it with questions, and troubleshoot scenarios visually through the camera instead of needing to convey the situation as a written question,” OpenAI explained in its website.