Generative AI has become a common way for young adults to look up information and seek advice. A Rutgers Health study now puts numbers on its use in a particularly sensitive area: substance use. Survey data from nearly 1,000 participants found that more than 32% of people ages 18 to 29 had used chatbots such as ChatGPT for substance-use information.
The study, published in the Journal of Medical Internet Research Formative Research, found that the most common reason was to learn about risks and consequences. About half of respondents who used AI for substance-related information said they did so at least monthly.
The result highlights a widening gap between how people seek health information and the safeguards available in conversational AI systems. Researchers said the pattern is concerning because chatbots can reproduce misinformation, while young adults are a group particularly affected by adverse effects of substance use.
AI is becoming part of health-information seeking
The Rutgers team examined how frequently young adults turn to generative AI for substance-related questions, the characteristics of people who do so, and what they think of the information the tools produce. Generative AI refers to systems that create synthetic content in response to prompts.
The work builds on survey data indicating that 66% of young adults use AI for activities that include searching for information and getting medical advice. But substance-use questions deserve separate attention. They can involve uncertainty about risks, consequences and behavior, and inaccurate information may carry immediate consequences.
In the Rutgers survey, roughly one in three participants reported using AI specifically for information related to substances and substance use. The researchers did not present the finding as evidence that every AI interaction is harmful. Instead, it shows that chatbot use in this context is already substantial enough to matter for public-health communication and for the design of AI systems.
Why the findings raise concern
Chatbots can generate fluent, confident-seeming responses. That presentation can make it difficult for a user to distinguish well-supported guidance from an error, a missing caveat or misinformation repeated from flawed material. In substance-use contexts, an incomplete answer could shape a decision about a risky behavior or delay a person from seeking more reliable support.
Lead author Mireia Triguero Roura, a postdoctoral researcher at Rutgers’ Center for Pharmacoepidemiology and Treatment Science, said that use of these tools could lead to misinformed or otherwise risky substance-use behavior among young people. The study’s concern is rooted both in the reliability limits of chatbots and in the vulnerability of the age group being studied.
The researchers found that respondents primarily used AI to seek information about risks and consequences. That is a meaningful distinction from casual experimentation with a new technology. It indicates that people may be treating chatbot answers as a source of guidance on health-related subjects that can be complex, personal and stigmatized.
Differences in who uses AI for drug-related questions
The study also identified differences among respondents. Black and male participants were more likely to report using AI for substance-use-related purposes. The researchers said those disparities could reflect differences in healthcare access, stigma, trust and health-information-seeking behavior.
That interpretation remains a question for further research rather than a conclusion about why any individual chooses a chatbot. Still, it points to an important equity issue. If some groups are more likely to rely on AI for drug-related questions, then weaknesses in the information produced by those systems may not be distributed evenly.
It also suggests that improving the safety of AI answers cannot be separated from improving access to trusted health information. Young adults may choose a chatbot because it is fast, private and available at any time. Those features can be especially attractive for questions people may hesitate to ask a clinician, family member or conventional search service.
What should follow
The Rutgers research does not prescribe a single solution, but it clarifies the need for closer attention from AI developers, health professionals and public-health institutions. For developers, the findings strengthen the case for careful handling of substance-related prompts and for making uncertainty and limits clear. For health organizations, they underline the need to make reliable information accessible in forms that match how young adults actually search.
For users, the most practical implication is that a chatbot should not be treated as a sole authority for substance-use information. An AI response may be a starting point for learning about a topic, but its claims should be checked against qualified clinical or public-health sources—especially when a decision could affect health or safety.
The study captures a broader shift in information behavior. Generative AI is no longer only a tool for writing, coding or general web-like questions. It is entering areas where trust, accuracy and context are central. The fact that about one-third of surveyed young adults have already used it for substance-use information makes that shift difficult to ignore.
As use becomes routine, the key question will not simply be whether people consult AI. It will be whether the systems and the surrounding health-information ecosystem can give them answers that are accurate, appropriately cautious and connected to dependable human support when it is needed.
This article is based on reporting by Medical Xpress. Read the original article.
Originally published on medicalxpress.com
