Twitter Bots Comprised Nearly Half of COVID-19 Discussion Accounts
A Carnegie Mellon University study found that bots comprised up to 60 percent of Twitter accounts participating in COVID-19 discussions. These bots were used to amplify calls to reopen the country, undermining the integrity of public discourse.
Bots may account for between 45 and 60% of Twitter accounts discussing covid-19.
Key facts
- What
- A Carnegie Mellon University study found that bots comprised up to 60 percent of Twitter accounts participating in COVID-19 discussions.
- Incident date
- May 21, 2020
- Who
- Failure mode
- Brand & Safety Incident
- AI surface
- Algorithmic Decision
- Severity
- High
What happened
Researchers from Carnegie Mellon University found that bots comprised between 45 and 60 percent of Twitter accounts discussing COVID-19. These bots were used to amplify specific messaging, particularly calls to reopen the United States. A majority of the most influential retweeters were likely bots.
What broke inside the model
- 01 · TriggerA user prompts the model in public view.
- 02 · Model stepThe model produces unsafe or off-brand output.
- 03 · Control gapNo filter holds the line before publish.
- 04 · FailureThe output goes public unchecked.
- 05 · ConsequenceA reputational or safety incident lands.
A contained signal crosses into output that goes public.
Twitter's automated bot detection systems failed to identify and mitigate large-scale orchestrated bot campaigns. This allowed bot accounts to maintain high visibility and manipulate public discourse.
What it cost
Sources
- PrimaryNearly Half of the Twitter Accounts Discussing 'Reopening America' May Be Botscmu.edu
- PressNearly half of Twitter accounts pushing to reopen America are botstechnologyreview.com
Cite this entry
https://failureindex.ai/failures/twitter-bots-comprised-nearly-half-covidAI Failure Index. "Twitter Bots Comprised Nearly Half of COVID-19 Discussion Accounts" (FI-0600). Realm Labs. https://failureindex.ai/failures/twitter-bots-comprised-nearly-half-covid (indexed Jun 22, 2026).Data fields CC-BY 4.0, prose citation permitted. Incident ID FI-0600. Full dataset at /data.
Note from Realm Labs, the Index steward
How Realm would have caught this
- Prism
- OmniGuard
- AI Detection & Response (AIDR)
Realm watches the model's internal state for the signature of unsafe or off-brand generation and can block or reroute the output before it becomes public, in real time rather than after it has been screenshotted.