When researchers submitted 20 ads with “blatant” voting misinformation, TikTok approved 90%, Facebook approved a “significant number”, and YouTube approved zero
Facebook’s approval of a significant share of the researchers’ submissions extends a documented pattern: a prior test of COVID-rule-breaking ads also found its screening system approving prohibited campaigns, while an earlier fake-senator ad test exposed identity-verification failures. The new comparison matters because TikTok, Facebook, and YouTube produced sharply different outcomes on the same voting-misinformation submissions.
YouTube’s zero approvals stand against its earlier history of carrying a misleading Biden-corruption ad, underscoring that ad-enforcement results can differ by test, policy area, and platform.
First-order effects
TikTok’s review process permitted 18 of the 20 test ads, making it the clearest immediate route for the researchers’ voting-misinformation material to clear paid-ad screening.
Facebook’s approval of a significant number of submissions leaves its ad safeguards exposed alongside TikTok’s, while YouTube’s rejection of all 20 separates its result in this test.
Second-order effects
The gap in approvals puts Facebook and TikTok under direct comparative pressure: YouTube’s outcome provides a concrete benchmark for screening the same set of submissions.
For actors seeking to distribute false voting claims through paid placements, platform choice becomes consequential when otherwise identical ads receive materially different review decisions.
Third-order effects
Repeated tests across Facebook’s COVID-policy, identity-verification, and voting-ad workflows point to ad enforcement as an operational weakness that can persist across abuse types rather than a one-off moderation lapse.
If comparative testing continues to show uneven approval rates, trust in platform ad-review systems will increasingly depend on demonstrated enforcement outcomes rather than stated policies.
The trend: Political-ad integrity is becoming a test of whether platforms can apply their published rules consistently in paid distribution channels.
TikTok fails to catch almost all misinformation in political ads - and yet it's increasingly the platform of choice for political campaigns. I shudder to think what this means for our democracies. https://twitter.com/...
@katieharbath It is interesting. Not a fan of CNN quoting the report saying Facebook approved a “significant number” without specifying. Is 1 significant? 5?
Narrator: We learned last month from a deposition of a Facebook witness that there was an entire category of revenue tracking TikTok ads that violated Facebook policies. Highly relevant to this @CNNBusiness report. https://twitter.com/... https://twitter.com/... https://twitter.c…
this is SO bad. these aren't random posts, nor are they are they minor / subtle violations — the companies are accepting paid ads lying about how & when to vote. https://twitter.com/...
YouTube and Facebook letting Brazil election disinformation spread, Global Witness says. The NGO produced a series of purposely misleading ads about the election — YouTube approved all of them to run and Facebook approved half. https://www.theguardian.com/ ...
In the experiments, ads submitted included false claims ton include “that voting days would be extended and that social media accounts could double as a means of voter verification. The ads also contained claims designed to discourage voter turnout.” https://twitter.com/...
TikTok and Facebook failed to detect ads with blatant election disinformation in an experiment by @cyber4democracy and @Global_Witness. These platforms must do better - rules that only exist on paper don't protect our democracy. #Midterms2022 https://cybersecurityfordemocracy.org…
‘You would think that they would have learned by now.’ Facebook and other social media still serve up election lies weeks before midterms https://fortune.com/...
“They don't think it matters because it is Brazil. Facebook and Google's shareholders are in the US, not here.” Social media giants spreading fake news ahead of run-off presidential vote, investigation reveals. Story by @adowniebrazil https://www.theguardian.com/ ...
It's not just @facebook. @Global_Witness tested @YouTube's ability to remove ads containing election disinformation before #Brazil's 2nd round. @YouTube approved 100% of our disinfo ads. And @Meta is STILL approving 50% #Eleiçoes2022 https://www.theguardian.com/ ...