MIT Announces Up to $53 Million Project to Develop Safer Weed Killers With AI

MIT announced Tuesday that its researchers will lead a project receiving up to $53 million from the Advanced Research Projects Agency for Health to develop alternatives to today's weed killers.

The project, called Precision Learning for On-target Weed control, or PLOW, will use generative artificial intelligence to identify chemical compounds and small interfering RNA that could control weeds while reducing risks to people and the environment.

Researchers say the work addresses two challenges: weeds are developing resistance to widely used herbicides, and the United States relies on imported ingredients used to make some of those products. The team aims to design candidates that are effective, affordable to produce and less likely to trigger resistance.

AI-generated candidates will move through laboratory screening before researchers test the most promising options in greenhouses. MIT says the first greenhouse tests of initial RNA candidates are expected by the end of the program's first year.

The team also plans to develop microbial systems that break down agricultural chemicals before they reach food, soil or waterways. The project includes researchers from MIT, the Broad Institute, several other universities and Robigo.

The work is research in development, not a newly approved herbicide. The funding is part of ARPA-H's STREAM program; the award began in September, and MIT detailed the project in its November 10 announcement.