Most people strongly support the federal government's reclassification of cannabis, according to a new study that used AI to analyze more than 40,000 comments in the public record.
The findings by researchers at Johns Hopkins University and the University of California San Diego, which come as the Trump administration last week reclassified state-licensed medical marijuana from a Schedule I drug to a less dangerous Schedule III, suggest the public would like to see even more reform.
"Americans showed up with their personal stories about therapy that helped them, businesses they built, consequences of cannabis use, and more," said first author Vijay M. Tiyyala, a Johns Hopkins research assistant.
The work is newly published in the journal Addiction.
In response to a 2024 proposal by the Drug Enforcement Administration to reschedule marijuana, 42,913 people submitted comments to the docket, which is part of the federal e-rulemaking process and how the public can engage with proposed regulations. It was the largest body of public input to date on federal cannabis policy.
The research team obtained every comment posted to the e-rulemaking portal during the 63-day comment window and used an AI-driven analysis validated against human review. Most commenters supported the change:
- 63.5% supported even more reform
- 28.9% supported Schedule III as proposed
- 6.7% opposed any change
Supporters of Schedule III included patients, healthcare providers, and business owners, often writing from personal experience. "This is a legit medication, it has saved my life and I'm in the medical field," one supporter wrote.
The analysis allowed the researchers to study what motivated people's opinions.
Supporters most often cited therapeutic benefits (56.7%), economic impacts on the cannabis industry and state revenues (27.8%), and the need for clearer federal regulation to ensure public safety (24.4 %).
"Rescheduling is a meaningful first step, but the public record shows Americans want federal policy to go even further," said senior author John W. Ayers , vice chief of innovation at UC San Diego School of Medicine.
Commenters who said Schedule III was insufficient often called for full descheduling, which would remove the substance from federal control. "Cannabis as a whole should be descheduled in order to preserve the thousands of businesses, patients and community members that rely on the ability to grow, extract and consume cannabis," one person wrote.
This group cited therapeutic benefits (37.8%), economic impacts (28.6%), and criminal justice reform (26.5%). Social-justice concerns about disproportionate incarceration in marginalized communities appeared in roughly one in five of their comments.
"Public enthusiasm for cannabis as medicine needs to be matched by federal and state investment in rigorous cannabis research," said co-author Ryan Vandrey , a cannabis researcher at Johns Hopkins.
Rescheduling cannabis to Schedule III will expand the type of research that can be conducted as well as the pool of scientists who can engage in research on the health effects of cannabis use, Vandrey added.
Almost all commenters, 92.4%, wanted cannabis off Schedule I. Comments reflected 14 different types of justifications, ranging from therapeutic benefits and economic impacts to criminal justice, medical research, and comparisons with alcohol and opioids.
"Many commenters have lived under state legalization for medical or adult use for years already. Their expectation of federal policy reflects that reality," said co-author Johannes Thrul , an addiction researcher at Johns Hopkins.
The group opposing any change wrote most often from a public health standpoint. "Moving cannabis into a category of drugs that have a low risk of dependence gives the public false information about this drug," one opponent wrote. Opponents cited public health risks (100%), addictiveness (71.4%), and potential harms to children and adolescents (57.1%), with parents concerned about youth access prominent among them.
The study demonstrated how AI-driven analyses can help the public shape government policy and make regulations more responsive to what Americans want.
"AI can revolutionize the way in which Americans engage in our democracy," said co-author Mark Dredze, a computer scientist at Johns Hopkins who directs the Data Science and AI Institute .
Authors included: Cerina Dubois and Clarissa Madar of Johns Hopkins.
