Reddit Answers Flagged for Dangerous Medical Advice; Moderators Call for AI Controls

Moderators across Reddit have raised alarms after the platform’s AI feature, Reddit Answers, reportedly suggested dangerous medical advice in response to users’ queries. One moderator said the tool recommended that people with chronic pain stop their prescriptions and take high-dose kratom — an unregulated substance that is illegal in some states — and even suggested heroin as a pain-relief option. Other concerning responses included potentially harmful guidance for neonatal fever.

Community moderators, especially from health-focused subreddits, say they currently have no way to disable Reddit Answers or hide problematic responses from view. That lack of moderation controls has amplified concerns about the AI tool’s suitability for sensitive topics.

• Reported recommendations included high-dose kratom and heroin for chronic pain.
• Moderators cannot turn off or flag Reddit Answers within affected subreddits.
• Reddit updated Reddit Answers to no longer display “Related Answers” on sensitive topics, per a company statement.

Reddit told 404 Media that the update ensures “Related Answers” to sensitive topics will no longer be displayed on the post detail (conversation) page. The spokesperson added Reddit Answers excludes content from private, quarantined, and NSFW communities, as well as some mature topics. However, moderators report the tool still lacks granular controls that would let community leaders decide when and how AI appears in their communities.

AI-generated content on social platforms has a history of producing inaccurate or dangerous suggestions, especially around medical and safety-related queries. While Reddit’s quick tweak may reduce the visibility of problematic suggestions in certain contexts, moderators argue that platform-level changes won’t replace fine-grained tools that let communities enforce their own standards.

For more context, see the original report on Engadget (link opens in a new tab).

If you’re a moderator and have noticed problematic Reddit Answers responses in your subreddit, consider documenting examples and reaching out to Reddit support. Community-level tools — such as the ability to disable AI features per-subreddit or flag specific AI responses — would make moderation more manageable and reduce the risk of harmful misinformation.

Discussion: Do you think platforms should allow AI to answer sensitive medical questions, or should AI be blocked from these topics until clearer safeguards and moderator controls exist?