The Verge: Genesis Mission selected 278 AI projects, while laboratories and universities must test their hypotheses
The Verge: Genesis Mission selected 278 AI projects, while laboratories and universities must test their hypotheses
On July 24, The Verge published an analysis of the first projects chosen by the US Genesis Mission, a government program that applies AI to science. Robert Hart connects the program to the White House's new plan and asks whether science has enough trained people and suitable facilities to test the hypotheses generated by machines.
On July 22, the US Department of Energy selected 278 projects under the Genesis Mission for grant negotiations. Teams from national laboratories, universities, companies, and nonprofit organizations will receive access to computing resources, models, and research software.
The White House plan published on July 21, “Science: The New Golden Age”, proposes relying on individual researchers, new organizations, and corporate partnerships. A grant gives a team access to computing resources and models, but researchers and laboratories remain responsible for testing the hypothesis. AI can propose a molecule, a material, a relationship in the data, or an experimental design. A researcher conducts a controlled experiment, a laboratory obtains a measurement, and another group repeats the experiment using new samples.
In The Verge's analysis, physicist Andreas Karch describes the risk: AI will produce many plausible options, while fewer people will have the expertise needed to evaluate them.
“AI may have a few important ideas buried under mountains of empty material, while far fewer people will have the expertise needed to tell one from the other.”
In biomedicine, a laboratory experiment is followed by a clinical study, which determines whether a person's condition changes. Each transition from a model to an experiment, replication, and clinical testing eliminates some hypotheses that initially sounded convincing.
The debate over the Genesis Mission concerns this entire sequence: computing resources accelerate the search for possible answers, while universities and laboratories train people to design experiments, verify results, and prepare the next generation of researchers.