• Online mental health content is ripe with misinformation that can lead to real-world harm, but traditional mitigation solutions often overlook the behavioral factors driving its spread.
• Prebunking and debunking interventions can help inoculate users against misinformation while combatting the cognitive biases that fuel the acceptance of falsehoods.
• AI content moderation and nudging are potent tools for mitigating the rapid sharing and virality of inaccurate (but highly engaging) mental health content.
• With the right tools and training, influencers are well-positioned to correct misinformation in online communities typically dominated by non-expert voices.

Most discussions concerning health misinformation focus on illnesses like COVID-19 or related interventions like vaccines. Mental health misinformation, though less well-studied, is just as concerning. With mental disorders estimated to account for 14% of the global disease burden, mental wellness has become a growing public health concern.³ As mental health disorders grow in prevalence, so do misconceptions—an issue that is too often reflected in the rampant spread of misinformation online.

Recent studies have uncovered shocking statistics about the reliability and accuracy of online mental health content. In one report exploring TikTok videos tagged with #mentalhealthadvice or #mentalhealthtips, 83% were found to be misleading, 31% contained scientifically inaccurate information, and 14% were potentially damaging—for example, suggesting that people take medication to treat a mental illness without consulting a doctor first.⁴ Even more concerning, these videos were viewed 25 million times and received over 3.5 million likes. In other research, 18% of videos about cognitive behavioral therapy (CBT) demonstrated negative attitudes about the treatment, with many erroneously claiming that CBT is ineffective or even harmful for specific populations—such as those with trauma or neurodevelopmental disorders—in which it has actually shown significant benefits.⁵

Needless to say, mental health misinformation can lead to several real-world problems. As an increasing number of users turn to online communities for mental health advice or support, inaccurate information risks pointing users toward harmful behaviors, such as stopping medication, avoiding professional help, or adopting ineffective or self-destructive coping mechanisms.⁶ 

Science has traditionally relied on the information deficit model to respond to these kinds of inaccuracies, assuming that belief in misinformation is rooted in a lack of access to facts.² The problem with this model is that it overlooks the important psychosocial factors that influence how we appreciate the truth. In short, facts alone don’t change minds. In the following sections, we’ll break down some of the key drivers behind the growing prevalence of misinformation and the leading interventions that could help reduce its influence.

Understanding the mental shortcuts that our brains take when processing information can help explain why inaccurate information often feels true. Cognitive biases make us more susceptible to falsehoods, especially when our emotions are running high, like when we’re confronted with mental health content that feels personally meaningful. Although there are likely several cognitive biases at play here, let’s explore a few to illustrate how they can contribute to widespread beliefs in misinformation.

Cognitive Biases Increase Our Susceptibility to Misinformation

Most misleading mental shortcuts are rooted in the brain’s tendency to seek the path of least resistance. Our hard-working noggins appreciate cognitive ease, so we’re often drawn to simple, emotionally compelling, and attention-grabbing stories over complex, evidence-based explanations. For example, “quick fix” cures for anxiety and depression offer easily digestible solutions, downplaying the importance of seeking professional help.

For the same reason, we often favor information that aligns with what we already know and believe. This tendency is known as the confirmation bias, which can cause us to search for and interpret information in a way that confirms our existing beliefs. As a result, people tend to avoid or diminish contradictory evidence.⁷ For example, someone who already suspects they have a generalized anxiety disorder may be prone to believing information from a video titled “¹⁰ Signs You Have Generalized Anxiety Disorder” while dismissing another video explaining how anxiety symptoms can vary and don't always fit the diagnostic criteria for a disorder.

We’re even more likely to believe inaccurate or misleading mental health content when we encounter the same information repeatedly. Thanks to the illusory truth effect, repeated exposure to misinformation can make that information seem more accurate and, thus, believable. This effect relies on familiarity, processing fluency, and understanding as signals of truth.² This can have a direct influence on our behavior. For example, one study found that participants who watched ADHD-related content more frequently were more likely to recommend inaccurate videos to others as educational content.⁸ Participants with a greater frequency of watching ADHD-related videos also gave higher accuracy and helpfulness scores to videos that were rated poorly by psychologists.

Mental Health Literacy Gaps Compound the Issue

Coupled with these cognitive biases, lower levels of mental health literacy can make users even more vulnerable to misinformation. For instance, a study on Italian online mental health communities found that users with low depression literacy were more likely to agree with misinformation provided by their peers.³ Interestingly, users who had higher levels of depression literacy were not more likely to believe misinformation the more they were exposed to it, suggesting that education is efficacious in combating the influence of the illusory truth effect on misinformation susceptibility. 

Equipping users with the tools to recognize misleading information is crucial for preventing false ideas from exploiting our cognitive biases and taking root. The best-known way to do this is a combination of prebunking (immunizing users to misinformation before it’s encountered) and debunking (correcting false information after it’s encountered). 

Prebunking to Inoculate Users Against Misinformation

Prebunking is an excellent first line of defense against misinformation. This proactive approach involves forewarning people of potential misinformation or preemptively refuting general claims that are currently circulating, essentially priming users to watch out for informational inaccuracies.² Typically, prebunking is broad in nature—think public service announcements telling people to watch for dramatic language in mental health content or banners on social media sites reminding users to be skeptical of one-size-fits-all mental health advice. 

How effective is this proactive intervention? Prebunking relies on inoculation theory, a sociological model that suggests people can develop a resistance to misleading messages by exposing them to weakened versions of this information. Much like how vaccines deliver a weakened dose of a virus, triggering the production of protective antibodies, prebunking does the same with ideas. 

Emerging research on the efficacy of inoculation against mental health misinformation suggests that this could be a valuable strategy for preventing harmful content from circulating online. For instance, one study found that preemptive corrective messages were effective at promoting resistance to misinformation linking mental illness and violence.²¹ Not only did these inoculation messages reduce the perceived credibility of misinformation, but they also increased the participants’ intentions to debunk stereotypes about mental health overall. Prebunking has also been found effective at immunizing people against misinformation related to other hot-button issues like climate change, immigration, conspiracy theories, and COVID-19.⁹ The efficacy of prebunking even seems to extend a kind of immunity umbrella over users, helping them build resistance to misleading persuasive techniques across various topics.² 

The main issue with prebunking is that it’s not always possible to predict the kind of misinformation people will be exposed to. While it’s best for helping people build a general defense against misinformation, debunking is essential for confronting issues with specific pieces of misinformation, like a viral video encouraging an untested depression treatment or an influencer sharing misleading facts about ADHD medication.

Debunking Misinformation to Reduce Its Impact on Behavior

Debunking is the reactive, defensive alternative to proactive prebunking. It involves targeting concrete instances of misinformation with specific counter-evidence. Say a licensed psychologist wants to post a video debunking a trending mental health claim that nail-biting is a sign of childhood trauma. An effective response would involve explaining why nail-biting is not a reliable or definitive indicator of mental health issues. Notice how it’s not enough just to say, “The claims presented in this video are untrue.” For debunking to be most effective, it must directly address the specific piece of information, explain why the original claim is misleading, provide a correction, and offer an alternative explanation grounded in evidence.² 

Since information is more easily consumed when it is simple, engaging, and personally relevant, framing debunking messages with personal stories or relatable narratives—instead of pure logic or complicated facts—can help make them more digestible, especially for those with lower mental health literacy. For example, this might mean sharing a real-life story about someone who has benefited from CBT or creating an engaging infographic breaking down how ADHD influences the brain. When debunking follows these best practices, it is generally effective at reducing belief in misinformation and mitigating its impact on behavior.²

While credible information is often unsensational, fact-based, and sometimes “dry,” misleading mental health content tends to be emotional, shocking, and attention-grabbing. These qualities increase the likelihood that misinformation is amplified by both human users and platform algorithms designed to maximize engagement. To understand how to combat this, let’s take a closer look at the algorithmic and psychosocial factors that contribute to the spread of misinformation.

Social Media Algorithms Reward Attention-Grabbing Content

Because social media platforms are designed to drive engagement, virality favors content that is emotional and novel—this content gets more views, likes, and shares, which keeps users on the platform. Regardless of its accuracy, platform algorithms push highly engaging content to users’ feeds, meaning that sensationalist stories often travel faster and reach more people than the truth. For instance, a recent study examining videos tagged #adhdtest—in which 92% of videos were found to be misleading—useful content had minimal engagement, gaining a measly 4% of total likes and 1% of total comments.¹⁰

Besides generally prioritizing engaging content across the board, algorithms also strive to provide individual users with personalized content. This means users see what they want to see, ultimately reinforcing what they already believe and limiting their access to conflicting perspectives. The result is echo chambers, or online environments where participants only encounter beliefs or opinions that coincide with their own.² Users stuck in echo chambers may also perceive a false consensus of their views, overestimating how many other people share their beliefs. Clearly, it’s not just engagement-driven algorithms that are to blame for the rapid spread of mental health misinformation. Humans also play a key role in fueling the fire.

Users Are Motivated to Share Content Regardless of Accuracy

There are a variety of psychosocial factors that encourage users to share information online, even if its accuracy is questionable. Many experts have examined this tendency through the framework of social identity theory, which proposes that people tend to form “ingroups” based on shared characteristics or beliefs.¹¹ Not only that, but people actually derive a part of their self-identity from these groups. To maintain this sense of identity, members of ingroups are motivated to project a positive image while at the same time differentiating themselves from “outgroups.” 

Sharing mental health content that’s consistent with ingroup views—whether this means criticizing mainstream therapy techniques, romanticizing burnout, or dismissing psychiatric medications as harmful—can affirm one’s membership within the group and even elevate their status. As a result, conflicting information rarely breaks into these circles, and when it does, it’s quickly rejected by group members. Mental health misinformation can spread rapidly among groups where member identities are rooted in skepticism about mainstream science. We saw this during COVID-19 when rare cases of functional neurological disorders (FND) following vaccination were misrepresented in viral social media videos as widespread risks, often taken as evidence of vaccine danger among anti-vax groups.¹² To address these challenges effectively, we must consider strategies that disrupt the amplification of misinformation while empowering users to critically evaluate content before sharing.

Countering the spread of mental health misinformation means working against both engagement-based algorithms and powerful human motivational factors. Unfortunately, traditional content moderation strategies—flagging posts and manually reviewing content for misinformation—struggle to keep up with the speed and scale of misinformation spread.¹³ For example, a recent content analysis found extremely high levels of misinformation in online mental health discussion spaces, even when these were moderated by mental health professionals.⁶ Existing moderation systems have also come under scrutiny for practicing censorship and limiting free speech, provoking reactance from some users. These issues highlight another excellent opportunity to combine technology and behavioral science. Rather than simply policing content, AI-powered moderation and proactive nudging can help change how people engage with it—before it goes viral.

Using AI to Identify Misinformation in Real Time

Advanced AI systems have the power to analyze patterns, language, and context more accurately than simple algorithmic systems and with far greater speed than humans.¹⁴ Researchers at Keele University recently developed an AI tool that can detect fake news with 99% accuracy, surpassing the performance of existing systems.¹⁵ One of the most significant benefits of adaptive AI is that it can understand nuance. Instead of automatically deleting or down-ranking content that contains certain keywords, it can add context, insert corrections to improve accuracy, or prime users to think critically before viewing emotionally charged content. For example, if a social media video suggests that positive thinking can cure depression, an AI moderation tool could display an automatic warning that the content may contain inaccurate information. A well-timed intervention like this encourages critical thinking while respecting users’ freedom to make their own decisions.

Nudging Users to Reconsider Sharing Misleading Content

Content warnings are fantastic examples of nudges—subtle cues designed to influence behavior without restricting choice. Nudges are effective at prompting users to consider accuracy before sharing potentially misleading content, reducing misinformation spread without threatening user autonomy or ridiculing people for making mistakes.² For example, a gentle accuracy nudge might be as simple as a statement saying, “Please consider the accuracy of this content before sharing it.” 

Nudges could also highlight the importance of sharing only true information, like, “It is widely accepted that spreading misinformation about mental health is dangerous.” Social proof nudges like these have the power to be even more effective, tapping into users’ desire to conform to social norms. One study used the phrase, “More than 80% of U.S. adults think it’s very important to only share accurate content online,” finding that it was very effective at reducing the number of false headlines shared by participants.¹⁶ Pairing nudges with AI-powered moderation is a great way to disrupt the sharing of misinformation while respecting user autonomy.

One final glaring issue with mental health content is the lack of experts in online spaces where this content circulates. More often than not, mental health information is posted and shared by users without training or expertise, leading to a lack of scientific accuracy in peer-to-peer discussions.³ For example, in a study evaluating the accuracy of information in Facebook communities for mental health, 26% of comments consisted of medically inaccurate information, and nearly 60% of threads contained at least one misinformed statement without any attempt at correction.⁶ Even moderators were not exempt from posting inaccurate information. In peer-led communities like these, misinformation that pops up in the comments of a thread—rather than the original post itself—has an even lower chance of being corrected. 

This authority gap creates environments where users begin to trust self-proclaimed mental health experts, influencers presenting non-expert opinions as fact, and commenters sharing emotionally charged anecdotal stories over the scientific consensus. Our tendency to put our trust in authority figures, also known as the authority bias, often causes us to defer to people with perceived authority—like influencers and armchair experts—regardless of their qualifications.

The Authority Gap Undermines Trust in Established Practices

The authority gap doesn’t just make it easier for false information to spread unchecked, but it also undermines trust in evidence-based information. Skepticism toward CBT is an excellent example, where a lack of experts in online communities has allowed misconceptions about the treatment to flourish. One exploratory study uncovered several common critiques about CBT in online videos claiming that it is ineffective, invalidating, or even harmful, especially for people with trauma, PTSD, neurodevelopmental disorders, and those facing systemic oppression.¹⁷ 

These reports are surprising, given that CBT treatment trials show very low rates of negative outcomes. In fact, CBT is one of the few treatments with robust empirical support for PTSD, yet someone relying on social media for non-expert information about the efficacy of CBT might conclude that this is up for debate.¹⁷ To address these issues, we need to bring credible expertise into online spaces.

The lack of experts in online communities presents a unique opportunity to mitigate the spread of misinformation: peer networks. Users already turn to peers in online communities for support and information, so leveraging these existing discussions and trusted relationships is the natural way to improve information accuracy. Part of the reason peer-to-peer corrections may be beneficial is that empathetic communication is more impactful at persuasion than wielding expertise.² There is also emerging evidence that corrections are more impactful when they come from people in our social circles rather than strangers.² 

Leveraging mental health content creators within groups—rather than outside experts—ensures messages align with the social norms and values of the group, reducing resistance and making corrections feel less threatening to social identities. For example, an influential mental health advocate sharing their positive experiences with anxiety medication may be more effective at promoting evidence-based treatment options to a “natural medicine” group than a clinical psychologist presenting impersonal facts and stats.

Of course, if we’re going to rely on wellness influencers as a reliable line of defense against misinformation, we have to adequately equip them to promote scientific accuracy. Fortunately, some researchers—and even a few social platforms—are already exploring this approach. In one recent experiment, researchers provided mental health content creators with digital toolkits containing evidence-based mental health information written in plain language and spanning several topics.¹⁸ They found that influencers who received the toolkits were significantly more likely to include evidence-based mental health content in their videos. Amazingly, TikTok videos featuring this accurate content received over half a million more views after the toolkit intervention, suggesting that accurate mental health content relayed by influencers can be incredibly engaging. A follow-up study even found that commenters exhibited better mental health literacy after viewing the videos.¹⁹ 

YouTube has implemented a similar approach with the launch of THE-IQ Creator Program in 2023, which allows eligible content creators to access workshops on best practices for video production, support from YouTube specialists, and funds to assist with content creation. THE-IQ aims to improve the accessibility of high-quality information and help viewers make better health decisions.²⁰ 

While efforts to educate content creators are a great step in the right direction, these methods largely rely on voluntary participation. Beyond education, ensuring the consistent quality of online mental health information requires the introduction of specific guidelines—and perhaps even some level of regulation—by both platforms and governments. At the very least, establishing clear standards for all mental health-related content can help ensure that anyone sharing this information knows how to do so responsibly. With guidance from mental health literacy programs, platform-provided toolkits, and standardized guidelines, content creators within online communities can play an impactful role in sharing evidence-based content and correcting inaccuracies among peers.

Each of the opportunities presented above introduces a novel approach to handling misinformation in online mental health content. However, the improper application of some of these potential solutions risks the same issues as traditional mitigation approaches. For example, overcorrecting users or outright censoring certain mental health content can erode trust in platforms, causing people to seek out alternative, often less reliable sources of information. Balancing free speech while controlling misinformation is always going to be tricky. Social media platforms—like Facebook, which recently ended its third-party fact-checking program—are still struggling to get this right. 

Another risk is that AI moderation and nudging strategies could be perceived as manipulative, making users less receptive to future corrections. For example, users who frequently encounter accuracy nudges before sharing content that they believe is accurate might start to dismiss these nudges altogether. This goes hand in hand with a risk of bias in interventions—who decides what mental health information is true or accurate? Some forms of misinformation may be easier to debunk or correct than others due to the availability of evidence-based information, which could create inconsistencies in how misinformation is addressed. While these innovative solutions show promise, they must be carefully implemented to avoid inciting pushback from users.

Combating the spread of mental health misinformation online is a long-term process that requires interdisciplinary collaboration from various stakeholders, from content creators and journalists to social media platforms and public health organizations. In the meantime, targeted interventions grounded in behavioral science can be valuable for reducing the acceptance of misinformation, mitigating misinformation sharing, and introducing effective peer-to-peer corrections in online communities.

Beyond focusing specifically on instances of online misinformation, this issue also demands broader educational efforts to boost mental health literacy and address negative stereotypes that perpetuate stigma, ensuring mental health discussions are accurate and productive. Thankfully, the efforts of international agencies like the World Health Organization (WHO) are helping to address inaccuracies in mental health content on a global scale. WHO has worked with social media policy departments around the world to produce guidelines for content creators and prevent the proliferation of false medical information on platforms like YouTube. Similarly, the Blended Intensive Programme in Bulgaria—taking place in 2025—offers university students and faculty from all over Europe the chance to learn critical skills in analyzing and responding to misinformation in a public health context.

As creators, researchers, social media companies, and global organizations collectively work to build user resilience to mental health misinformation, The Decision Lab is eager to collaborate with stakeholders to develop and implement interventions leveraging behavioral science. Our expertise in this area extends from behavioral machine learning and digital product design to the development of intentional policies to improve global well-being. Contact us today to help us foster a culture of truth around digital mental health content.

The Misinformation Mitigation Toolbox: Dismantling the Digital Deceit

Online misinformation is a significant problem extending beyond mental health content, misleading users across other high-stakes areas from politics to climate change. In this article, we present a toolbox of interventions to help combat misinformation, improve information literacy, and promote critical thinking in online environments.

A New SPIN on Misinformation

Effectively evaluating and moderating content first requires understanding the various types of misinformation that spread online. TDL developed a taxonomy called Sorting Potentially Inaccurate Narratives (SPIN) to facilitate this process. Our goal with this taxonomy is to support the design of targeted interventions that help online users identify and avoid misinformation.

1. Vosoughi, S., Roy, D., & Aral, S. (2018). The spread of true and false news online. Science, 359(6380), 1146-1151. https://doi.org/10.1126/science.aap9559 
2. Ecker, U. K., Lewandowsky, S., Cook, J., Schmid, P., Fazio, L. K., Brashier, N., ... & Amazeen, M. A. (2022). The psychological drivers of misinformation belief and its resistance to correction. Nature Reviews Psychology, 1(1), 13-29. https://doi.org/10.1038/s44159-021-00006-y 
3. Bizzotto, N., de Bruijn, G. J., & Schulz, P. J. (2023). Buffering against exposure to mental health misinformation in online communities on Facebook: the interplay of depression literacy and expert moderation. BMC public health, 23(1), 1577. https://doi.org/10.1186/s12889-023-16404-1
4. PlushCare. (2022). How accurate is mental health advice on TikTok? https://plushcare.com/blog/tiktok-mental-health/ 
5. Starvaggi, I., Dierckman, C., & Lorenzo-Luaces, L. (2024). Mental health misinformation on social media: Review and future directions. Current Opinion in Psychology, 56, 101738. https://doi.org/10.1016/j.copsyc.2023.101738
6. Bizzotto, N., Schulz, P. J., & de Bruijn, G. J. (2023). The “loci” of misinformation and its correction in peer-and expert-led online communities for mental health: Content analysis. Journal of medical Internet research, 25(1), e44656. https://doi.org/10.2196/44656 
7. Piksa, M., Noworyta, K., Gundersen, A., Kunst, J., Morzy, M., Piasecki, J., & Rygula, R. (2024). The impact of confirmation bias awareness on mitigating susceptibility to misinformation. Frontiers in public health, 12, 1414864. https://doi.org/10.3389/fpubh.2024.1414864 
8. Karasavva, V., Miller, C., Groves, N., Montiel, A., Canu, W., & Mikami, A. (2025). A double-edged hashtag: Evaluation of# ADHD-related TikTok content and its associations with perceptions of ADHD. PloS one, 20(3), e0319335. https://doi.org/10.1371/journal.pone.0319335 
9. Pilditch, T. D., Roozenbeek, J., Madsen, J. K., & van der Linden, S. (2022). Psychological inoculation can reduce susceptibility to misinformation in large rational agent networks. Royal Society open science, 9(8), 211953. https://doi.org/10.1098/rsos.211953
10. Verma, S., & Sinha, S. K. (2025). How evidence-based is the “hashtag ADHD test”(# adhdtest). A cross-sectional content analysis of TikTok videos on attention-deficit/hyperactivity disorder (ADHD) screening. Australasian Psychiatry, 33(1), 82-88. https://doi.org/10.1177/10398562241291956 
11. Craig, K., & Sadovykh, V. (2022). Perceived social media bias, social identity threat, and conspiracy theory ideation during the COVID-19 pandemic. Proceedings of the 55th Hawaii International Conference on System Sciences, Hawaii, United States. http://dx.doi.org/10.24251/HICSS.2022.726 
12. Butler, M., Coebergh, J., Safavi, F., Carson, A., Hallett, M., Michael, B., Pollak, T. A., Solomon, T., Stone, J., Nicholson, T. R., & Coronerve Studies Group (2021). Functional Neurological Disorder After SARS-CoV-2 Vaccines: Two Case Reports and Discussion of Potential Public Health Implications. The Journal of neuropsychiatry and clinical neurosciences, 33(4), 345–348. https://doi.org/10.1176/appi.neuropsych.21050116 
13. Baker, S. A., Wade, M., & Walsh, M. J. (2020). The challenges of responding to misinformation during a pandemic: content moderation and the limitations of the concept of harm. Media International Australia, 177(1), 103–107. https://doi.org/10.1177/1329878X20951301
14. Li, C., & Callegari, A. (2024, June 14). Stopping AI disinformation: Protecting truth in the digital world. World Economic Forum. https://www.weforum.org/stories/2024/06/ai-combat-online-misinformation-disinformation/
15. Asowo, P., Lal, S., & Ani, U. D. (2024, November). An Ensemble Modelling of Feature Engineering and Predictions for Enhanced Fake News Detection. In International Conference on Innovative Techniques and Applications of Artificial Intelligence (pp. 225-231). Cham: Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-77918-3_16 
16. Butler, L. H., Prike, T., & Ecker, U. K. (2024). Nudge-based misinformation interventions are effective in information environments with low misinformation prevalence. Scientific Reports, 14(1), 1-12. https://doi.org/10.1038/s41598-024-62286-7
17. Lorenzo-Luaces, L., Dierckman, C., & Adams, S. (2023). Attitudes and (Mis)information About Cognitive Behavioral Therapy on TikTok: An Analysis of Video Content. Journal of medical Internet research, 25, e45571. https://doi.org/10.2196/45571
18. Motta, M., Liu, Y., & Yarnell, A. (2024). “Influencing the influencers:” a field experimental approach to promoting effective mental health communication on TikTok. Scientific Reports, 14, 5864. https://doi.org/10.1038/s41598-024-56578-1 
19. Murphy-Reuter, B. (2025, January 16). Social media is the new public health frontline. Let’s treat it that way. Harvard Public Health Magazine. https://harvardpublichealth.org/tech-innovation/to-combat-misinformation-social-influencers-need-the-right-tools/
20. Graham, G. (2023, September 7). Expanding equitable access to health information on YouTube. YouTube Official Blog. https://blog.youtube/news-and-events/expanding-equitable-access-to-health-information-on-youtube/
21. Zhang, N. (2021). Inoculating the Public Against Misinformation: Testing the Effectiveness of" Pre-Bunking" Techniques in the Context of Mental Illness and Violence (Doctoral dissertation, University of South Carolina). https://scholarcommons.sc.edu/etd/6432