Social media and psychosis: friend, foe, or something in between?
Social media is now stitched into everyday life. For many people, it’s where friendships are maintained, news is consumed, and support is found. For others, it’s a source of stress, comparison, or misinformation. Research on how social media affects mental health has been inconclusive, and researchers still debate its effect on youth and adults. This discussion also sits against the backdrop of Meta’s recent nearly $18 billion settlement over allegations that its platforms contributed to psychological harms among young users. A reminder that concerns about social media effects on mental health are not abstract and continue to shape public debate, regulation, and clinical conversations.
When it concerns people living with schizophrenia or other psychotic disorders, this debate becomes even more layered. Some describe social media as a lifeline that helps them stay connected when symptoms make face‑to‑face interactions difficult. Others say it can heighten anxiety or make them feel vulnerable. Despite this, research on psychosis has focused almost entirely on the offline world. We still know surprisingly little about how this online engagement relates to the symptoms, functioning, or well-being of individuals with schizophrenia and other psychotic disorders.
This new systematic review by Maddalena Cipriani and colleagues from Harvard Medical School tries to shed light on this issue (2026). This study summarised evidence from 14 studies published since 2010 to explore whether social media helps, harms, or has no meaningful effect on people with schizophrenia and other psychotic disorders regarding their symptom severity. It’s a timely question: clinicians, carers, and service users regularly ask whether social media is safe, whether it worsens symptoms such as paranoia, or whether it can support recovery.
The review offers a nuanced picture, one that challenges assumptions and highlights the importance of how people use social media, not just how much.
Methods
This study is a systematic review; a type of research that gathers all the quality studies on a given topic (i.e., social media use and psychotic symptoms), checks their quality, and summarises what the overall evidence shows. To do so, they use specific methods to avoid bias and omissions and to ensure that their study can be replicated by another team of independent researchers. The authors followed up-to-date and important guidelines for their work (PRISMA) and preregistered their protocol (PROSPERO). They searched PubMed, Embase, and Medline for studies examining any association between social media use and mental health symptoms in people with schizophrenia or other psychotic disorders. They also assessed the risk of bias in each study.
In their review, they found 14 studies that met their eligibility criteria. Among them, nine studies were cross‑sectional (studies conducted at only one time point) and three longitudinal. Most studies were quantitative (done by analysing metrics), but one was qualitative (done by analysing participants’ interviews). One study used mixed methods. This broad inclusion strategy allowed the authors to capture a wide range of experiences and outcomes.
Results
The 14 studies amounted to 1,957 participants in total. Samples ranged from adolescents and young adults to middle-aged adults. Eight studies recruited participants from clinical settings such as outpatient clinics and early psychosis services. The other six recruited participants from online adverts and social media platforms. In these studies, “diagnoses could be self-reported or clinically established”. What follows are key findings of this review.
Negative symptoms and functioning
Psychotic negative symptoms are the aspects of schizophrenia where people lose or have reduced abilities, such as having less motivation, speaking less, showing fewer emotions, or withdrawing from social contact. Five cross-sectional studies found a consistent pattern: more social media use was linked with fewer negative symptoms and better psychosocial functioning. Social media use was also associated with higher psychosocial functioning and quality of life. It is important to note that this doesn’t prove that social media improves symptoms; far from it. It could be that people with fewer negative symptoms may simply be more active, functional, and capable of using social media. In other words, symptom severity could be both a cause and a consequence.
Positive symptoms and paranoia
Positive symptoms are experiences that add unusual perceptions or beliefs, such as hallucinations, delusions (e.g., paranoia), or disorganized thinking. Eight studies examined positive symptoms, particularly paranoia. The review summarises well: “Studies consistently find no significant association between social media use and the severity of positive symptoms”. However, among those studies, two longitudinal studies found small reductions in paranoia after social media use. For example, Paquin et al. (2024a) reported a ~10% decrease in paranoia, but how people used social media mattered. Berry and colleagues found that passive browsing predicted reduced paranoia, whereas “venting” on social media predicted lower mood.
Mood and problematic use
Findings were mixed. Some studies linked lower social media use with more depressive symptoms; others found higher use associated with more anxiety. The clearest pattern was that quality of connection mattered. Feeling “closer to others” online was associated with lower depression. Using social media in a compulsive or unhealthy way (problematic use) was linked to worse well-being. In one study, researchers found that youth at higher risk of developing psychosis experienced greater emotional distress associated with internet addiction. Another study found that problematic social media use was connected to poorer quality of life, mainly because it disrupted sleep and increased feelings of shame or self‑stigma.
Conclusions
- The authors conclude that social media is not inherently harmful for people with psychosis.
- Across the studies, it did not worsen psychotic symptoms and was often linked with better psychosocial functioning and even short‑term reductions in paranoia.
- The authors emphasise that how people use social media matters more than how much time they spend online.
- At the same time, they caution that most available evidence is limited by the cross‑sectional designs of the majority of studies and a moderate to high risk of bias the authors found in their analysis.
- Stronger longitudinal and experimental studies are needed to clarify whether social media use influences symptoms or simply reflects underlying changes in well-being.
Strengths and limitations
This review makes an important contribution by synthesising scattered and inconsistent literature. The authors believe it is the first systematic attempt to map how social media relates to symptoms and functioning in people living with psychosis. The inclusion of qualitative studies adds depth, and the authors’ attention to risk of bias is welcome. However, several limitations shape how confidently we can interpret the findings, and this is where critique becomes essential.
First, study quality was generally low for studies included in this systematic review. Most included studies were cross‑sectional, making it impossible to determine causes and consequences. Therefore, it is possible that the studies were reporting not on how social media shapes individual psychotic experiences, but rather on the characteristics of individuals who use social media more frequently. It is not surprising that individuals with less debilitating psychotic experiences are more often connected than their counterparts. Particularly, negative symptoms that have been associated with lower social media use include, among other things, loss of motivation, pleasure, and social interest. All elements necessary for engagement on social media. Another issue in study quality is that the risk of bias was moderate to high in nearly all quantitative studies. For example, many relied on non‑validated questionnaires, raising concerns about measurement accuracy. I should clarify that the authors acknowledge all those limitations in their paper, and the lack of quality evidence does not reflect the quality of this review.
Second, this is a high‑quality review, but I do have some concerns about how certain findings are presented. For example, the authors suggest that some studies point to a “potential protective effect” of social media on paranoia. This claim, which comes from two studies that capture short-term fluctuations in symptoms, should be treated very cautiously. A true protective effect would be reflected in long-term symptom improvement, not temporary symptom fluctuations. In addition, one of these two studies found no association between overall social media use and paranoia. Although seeing the newsfeed was linked to reduced paranoia, many other behaviors – such as posting about feelings – were associated with higher levels of paranoia. Taken together, it is difficult to interpret these findings as evidence of a protective effect. The review also does not engage with related research on psychotic‑like experiences in non‑clinical populations, where there are studies that, for example, have found that young adults who, on average, report higher levels of social media also report more psychotic‑like symptoms (Paquin et al., 2024b). This is not strong evidence of a negative effect, but does challenge the assumption that social media protects against psychotic symptoms.
The authors rightly highlight that how people use social media may matter more than how much they use it. While this aligns with broader digital well-being research, I think the framing risks placing too much responsibility on individual behaviour and too little on platform design. Behaviors linked to poorer well-being, such as doomscrolling or engaging with highly emotional content, are not accidental; they are often encouraged by design features intended to maximise engagement.
Given these limitations, I find that the review casts a somewhat optimistic light on social media that may not be fully supported by the evidence or reality itself. The authors acknowledge many of these constraints, but the overall framing still leans toward social media being beneficial for people with psychosis. A more cautious interpretation may be warranted. Nonetheless, despite some concerns about how certain issues are framed, this remains a high‑quality systematic review that brings together important evidence on how social media relates to psychotic symptoms among people living with mental health difficulties.
Implications for practice
Clinicians and people concerned with psychosis often worry that social media worsens paranoia or hallucinations. Current evidence presented in this systematic review does not support this. Instead, evidence suggests that problematic or addictive patterns were consistently linked to poorer mental health outcomes.
This is the part where most researchers will write that clinicians should avoid blanket discouragement and instead explore how patients use social media. Researchers would also add that digital literacy interventions encouraging healthier engagement could help people navigate online spaces more safely. Indeed, those are good recommendations. However, I believe that most of the responsibility for the mental health consequences of social media use lies with the companies that profit from those platforms and not with their users or clinicians. As such, this is a matter of public policy. Social media companies should have the incentives to not exploit users’ fragility, to not use exploitative persuasive design features that maximise engagement at all costs, and to provide a safe environment for everyone overall. The target for interventions should be design features and content presented on social media, not the behavior of each user, as is usually suggested.
Finally, I would be cautious about concluding that social media is safe, or even protective, for people living with psychosis based solely on this review. While the current review frames social media this way, another recent systematic review offers a more cautionary counterpoint (Yang & Crespi, 2025). They found that social media can heighten feelings of surveillance and self‑consciousness, potentially intensifying paranoia, and may even reinforce psychotic experiences by weakening cues needed for reality testing and by algorithmically feeding content that aligns with existing delusional beliefs. These findings suggest that social media may pose meaningful risks for individuals vulnerable to psychosis and that potential harm should be considered alongside the more optimistic interpretations offered in the current review. Content generated by artificial intelligence may now pose even greater risks.
My own experience also leads me to a more cautious stance. I have seen people with psychotic difficulties become deeply immersed in social media in ways that appeared to intensify delusional thinking and fuel a downward spiral. As a psychologist and researcher studying digital media, I would not recommend high levels of social media use for individuals living with psychosis, especially when there is a history of problematic or compulsive engagement.
Statement of interests
Gabriel Arantes Tiraboschi has no conflict of interest to declare. He used the Quillbot app to proofread and improve the readability of his text.
Edited by
Dr Simon Bradstreet.
Links
Primary paper
Cipriani M, Notsu H, Vaidya S, Castillo J, Hau C, Li S, Xia W, Yacoub D, Zhou SZ, Birnbaum ML, Torous J. The Relationship between Social Media and Schizophrenia and Other Psychotic Disorders: A Systematic Review. Schizophr Bull Open. 2026 May 27;7(1) 10.1093/schizbullopen/sgag023
Other references
Berry N, Emsley R, Lobban F, Bucci S. (2018) Social media and its relationship with mood, self-esteem and paranoia in psychosis. Acta Psychiatr Scand 138:558–570.
Paquin, V., Ackerman, R. A., Depp, C. A., Moore, R. C., Harvey, P. D., & Pinkham, A. E. (2024a). Media Use and Its Associations With Paranoia in Schizophrenia and Bipolar Disorder: Ecological Momentary Assessment. JMIR Mental Health, 11(1), e59198.
Paquin, V., Philippe, F. L., Shannon, H., Guimond, S., Ouellet-Morin, I., & Geoffroy, M.-C. (2024b). Associations between digital media use and psychotic experiences in young adults of Quebec, Canada: A longitudinal study. Social Psychiatry and Psychiatric Epidemiology, 59(1), 65–75.
Yang N, Crespi B. I tweet, therefore I am: a systematic review on social media use and disorders of the social brain. BMC Psychiatry. 2025 Feb 3;25(1):95.
