Facebook's AI fails to label misinformation variations ahead of US election
Avaaz's investigation found that Facebook's AI detection system failed to label 738 posts containing variations of fact-checked misinformation, allowing them to accumulate 142 million views. The posts included false claims about presidential candidates and mail-in ballots. Despite Facebook's claims that its AI can recognize near-duplicates, simple tweaks to content circumvented the system. Avaaz provided evidence to Facebook, but only 4% of the flagged posts were subsequently labeled.
- Date it happened
- 2020-10-09
- Organisation involved
This incident was imported from secure.avaaz.org. Our additions to it — the structured fields, the translation, the checks against other reports — are published under the same licence.
This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.