Custodes Futurifor the advancement of humanity

Misinformation

Misinformation is a barrier to human progress within the technological factors. The evidence for it, and the methods proposed to address it, follow.

Vosoughi, Roy, and Aral (2018) studied about 126,000 verified true and false news stories shared on Twitter from 2006 to 2017.

Evidence

  • Reach. False news spread significantly farther, faster, deeper, and more broadly than true news in every category.
  • Topic. The effect was strongest for political news.
  • Novelty. False news was more novel, which may explain why people shared it.
  • Human sharing. Automated accounts spread true and false news at the same rate, so the difference came from human sharing.
  • Effect on vaccination intent. A randomized trial included 8,001 online respondents in the UK and USA in September 2020. In the trial, recent COVID-19 vaccine misinformation, compared with factual information, lowered the share who would "definitely" accept a vaccine. The drop was 6.2 percentage points in the UK and 6.4 points in the USA. Scientific-sounding misinformation was more strongly associated with declines. The outcome was stated intent, not actual vaccination. The authors note the setting is unlikely to represent real social media exposure (Loomba et al., 2021).
  • Misleading true stories matter more than flagged falsehoods. A study combined experiments with 18,725 participants, crowdsourcing, and machine learning. With these, it estimated the effect of 13,206 vaccine-related links on the vaccination intentions of about 233 million US Facebook users. Flagged misinformation lowered intentions more per view, but few people saw it. Some content was not flagged yet encouraged vaccine skepticism, such as stories about rare deaths after vaccination. The study estimated that this content had an impact 46 times greater. The total effect is a model estimate built on survey experiments, not measured vaccination (Allen, Watts, and Rand, 2024).
  • Ranked a top global risk. The World Economic Forum's Global Risks Report 2026 is based on a survey of over 1,300 experts. In it, misinformation and disinformation ranked second among risks over the next two years and fourth over ten years. This measures expert concern, not demonstrated harm. Some researchers argue fears exceed the evidence. They say people largely avoid misinformation and assume others are more gullible than they are (Altay and Acerbi, 2023) (WEF, 2026).

Why it matters: When falsehoods spread faster than facts, public decisions on health, science, and policy suffer. This strengthens the identity-protective thinking described in the cultural factor.

What can be done

Actions rated on the strength-rating scale (A strong and replicated, B solid but limited, C weak or debated, D contested or failed), applied to the specific claim made.

  • Publish professional fact-checks and deliver them repeatedly (A for reducing belief soon after exposure, B for lasting effects). Simultaneous experiments in Argentina, Nigeria, South Africa, and the United Kingdom tested 22 fact-checks. In these experiments, fact-checks reduced belief in misinformation by at least 0.59 points on a 5-point scale. Most effects were still apparent more than 2 weeks later (Porter and Wood, 2021). In the United States, Great Britain, and Canada, fact-checks reduced COVID-19 misperceptions, especially among the people most vulnerable to them. But the reductions did not persist over time, even after repeated exposure (Carey et al., 2022). In South Africa, people were encouraged to receive fact-checks every two weeks by WhatsApp text or podcast for six months. This improved their ability to spot new political and health misinformation, especially when they were paid to read them. The encouragement had limited effects on news choices or checking behavior (Bowles et al., 2025). See "Cognitive dissonance and motivated reasoning" in the psychological factors for evidence that corrections do not backfire.
  • Media literacy programs with several sessions (B). A meta-analysis of 49 experimental studies with 81,155 participants found that media literacy interventions improved resilience to misinformation (d = 0.60, a medium effect). It found that the interventions reduced belief in misinformation (d = 0.27), improved the ability to tell true from false (d = 0.76), and reduced sharing (d = 1.04). Effects were stronger when there were multiple sessions (Huang, Jia, and Yu, 2026). In a randomized survey experiment, a brief media literacy intervention helped people tell false from correct information more generally. This held both immediately and about two weeks later. In the same experiment, fact-checking mainly affected the specific false story it addressed (Berger et al., 2025). Most studies measure outcomes in surveys soon after the intervention. Lateral reading is covered in "Cognitive dissonance and motivated reasoning" in the psychological factors. Short media literacy tips are covered in the proposed methods below.
  • Removing accounts that repeatedly spread misinformation (C). After the January 6, 2021, attack on the U.S. Capitol, Twitter suddenly removed 70,000 accounts trafficking in misinformation. A panel of more than 500,000 active users then showed less circulation of misinformation by those users and by their followers. Many such accounts that were not removed left the platform. But the size of the effect could not be identified, because the removal coincided with the events of that day (McCabe et al., 2024).
  • Not shown to work: one-hour in-person media literacy training in India (B). A field experiment with 1,224 people took place during India's 2019 elections. It gave training in reverse image searching and using a fact-checking website. The training did not significantly improve the ability to identify fake news. Supporters of the ruling party became significantly less able to identify fake news that matched their views (Badrinathan, 2021).

Proposed and experimental methods

Methods that are proposed, under trial, approved in some places, or tried and then failed. Each shows a stage label and an evidence rating. A stage label shows how far a method has progressed, not whether it works. The stage labels are explained on the technological factor page.

  • Fact-checker warning labels (Large trial, B). Platforms attach a warning from a professional fact-checker to false posts. Across 21 experiments with 14,133 participants, labels reduced belief in false headlines by 27.6% and sharing by 24.7%. Among those most distrusting of fact-checkers, labels still reduced belief by 12.9% and sharing by 16.7% (Martel and Rand, 2024). The experiments were online surveys rather than live platforms.
  • Short media literacy tips (Large trial, B). The study used preregistered survey experiments modeled on Facebook's tips for spotting false news. In a nationally representative U.S. sample, the tips improved people's ability to tell mainstream from false headlines by 26.5%. In a highly educated online sample in India, the gain was 17.5%. The U.S. effect was still measurable several weeks later. There was no effect in a largely rural area of northern India with low social media use (Guess et al., 2020). A correction to the article was published in 2023.
  • Platform regulation through the EU Digital Services Act (Approved but not scaled, C). The regulation was adopted on October 19, 2022 (European Parliament and Council, 2022). It requires platforms with over 45 million monthly EU users to analyze and reduce systemic risks, including to electoral processes and public health. It allows fines of up to 6% of global annual turnover. The Commission named the first 17 very large platforms and 2 search engines on April 25, 2023 (European Commission, n.d.). No study has yet measured the regulation's effect on misinformation.
  • Cross-references. Accuracy prompts and prebunking games and videos, with their replication disputes, are covered in "Cognitive dissonance and motivated reasoning" in the psychological factors. That section also covers Community Notes and feed reranking.
  • Reversed: third-party fact-checking at Meta in the U.S. (Failed or reversed, B). Meta said in January 2025 that it would replace U.S. third-party fact-checking with a crowd-sourced notes system (Associated Press, 2025). Meta made this choice even though the warning labels that fact-checks produce had reduced belief in and sharing of false posts in experiments (see the first entry in "Misinformation" in the technological factors).

Sources cited on this page

  1. Vosoughi, S., Roy, D., & Aral, S. (2018). The spread of true and false news online. Science, 359(6380), 1146-1151. DOI B Moderate
  2. Allen, J., Watts, D. J., & Rand, D. G. (2024). Quantifying the impact of misinformation and vaccine-skeptical content on Facebook. Science, 384(6699), eadk3451. DOI B Moderate: for direction, C for the size of the population effect
  3. Loomba, S., de Figueiredo, A., Piatek, S. J., de Graaf, K., & Larson, H. J. (2021). Measuring the impact of COVID-19 vaccine misinformation on vaccination intent in the UK and USA. Nature Human Behaviour, 5(3), 337-348. DOI B Moderate
  4. World Economic Forum (2026). Global Risks Report 2026 B Moderate: as a measure of expert concern, C as evidence of harm
  5. Guess, A. M., Lerner, M., Lyons, B., Montgomery, J. M., Nyhan, B., Reifler, J., et al. (2020). A digital media literacy intervention increases discernment between mainstream and false news in the United States and India. Proceedings of the National Academy of Sciences, 117(27), 15536-15545. link B Moderate
  6. Martel, C., & Rand, D. G. (2024). Fact-checker warning labels are effective even for those who distrust fact-checkers. Nature Human Behaviour, 8(10), 1957-1967. link B Moderate
  7. Ma, Y., Fairlie, R., Loyalka, P., & Rozelle, S. (2024). Isolating the "tech" from EdTech: Experimental evidence on computer-assisted learning in China. Economic Development and Cultural Change, 72(4), 1923-1962. link B Moderate
  8. Badrinathan, S. (2021). Educative interventions to combat misinformation: Evidence from a field experiment in India. American Political Science Review, 115(4), 1325-1341. link B Moderate
  9. Berger, L. M., Kerkhof, A., Mindl, F., & Münster, J. (2025). Debunking "fake news" on social media: Immediate and short-term effects of fact-checking and media literacy interventions. Journal of Public Economics, 245, 105345. link B Moderate
  10. Bowles, J., Croke, K., Larreguy, H., Liu, S., & Marshall, J. (2025). Sustaining exposure to fact-checks: Misinformation discernment, media consumption, and its political implications. American Political Science Review, 119(4), 1864-1887. link B Moderate
  11. Carey, J. M., Guess, A. M., Loewen, P. J., Merkley, E., Nyhan, B., Phillips, J. B., & Reifler, J. (2022). The ephemeral effects of fact-checks on COVID-19 misperceptions in the United States, Great Britain and Canada. Nature Human Behaviour, 6(2), 236-243. link B Moderate
  12. Huang, G., Jia, W., & Yu, W. (2026). Media literacy interventions improve resilience to misinformation: A meta-analytic investigation of overall effect and moderating factors. Communication Research, 53(7), 899-926. link B Moderate
  13. McCabe, S. D., Ferrari, D., Green, J., Lazer, D. M. J., & Esterling, K. M. (2024). Post-January 6th deplatforming reduced the reach of misinformation on Twitter. Nature, 630(8015), 132-140. link C Limited
  14. Porter, E., & Wood, T. J. (2021). The global effectiveness of fact-checking: Evidence from simultaneous experiments in Argentina, Nigeria, South Africa, and the United Kingdom. Proceedings of the National Academy of Sciences, 118(37), e2104235118. link A Strong

Every source for this factor is listed on the technological factor page.