Google's Jigsaw and YouTube to test video ads in Europe to “pre-bunk” anti-refugee sentiment, after a study found such ads raised awareness about misinformation
It also follows YouTube's 2017 brand-safety response, when Google expanded controls after ads appeared alongside extremist material in the UK. The related coverage frames the company’s misinformation work as part of a longer attempt to manage harms associated with its recommendation and advertising systems.
First-order effects
Jigsaw and YouTube will use European video-ad placements as a measured intervention intended to raise viewers’ awareness of misinformation before anti-refugee narratives take hold.
The test gives Jigsaw an additional distribution mechanism beyond the search-triggered Redirect Method, with YouTube supplying the audience reach and video inventory.
Second-order effects
YouTube’s trust-and-safety work shifts from responding to particular searches or harmful videos toward testing proactive messages delivered through its ad systems.
The reported awareness gains create an internal basis for Jigsaw to evaluate whether similar video campaigns warrant expansion, including the later Germany-focused prebunking test described in related coverage.
Third-order effects
If repeated tests show durable results, large platforms may treat ad targeting and creative testing as standing tools for misinformation prevention, alongside content rules and search interventions.
That approach would further blur YouTube’s separation between advertising infrastructure and safety operations, making campaign design and measurement part of platform-governance practice.
The trend: Platform safety efforts are moving from reactive removal and search redirection toward proactively distributing misinformation-resilience messages through advertising systems.
“It worked for conservatives and liberals. It worked for people who are already conspiratorial or not. It even worked for people who are very receptive to bullshit” (on the emerging power of prebunking in tackling misinfo) https://www.protocol.com/...
The study was ecologically realistic: there was almost no daylight between the study & how one would actually run a prebunking campaign on social media. View rate (how many people who were shown the ad actually watched it) was 19%. Total cost per ad view was $0.05 8/12
Importantly, these effects were robust across the political spectrum and a wide range of covariates. Effect sizes for manipulation technique recognition ranged between d = 0.28 and d = 0.68, very substantial for such short (+- 1.5 min) interventions 4/12
🚨New paper in @ScienceAdvances! In 7 studies (N=29,096), including an ecologically realistic field study on @YouTube, we find that #prebunking videos confer strong resistance against 5 manipulation techniques common in #misinformation https://doi.org/... 1/12 https://twitter.com/…
Across 6 lab studies (N = 6,464), we show that watching a prebunking video 1) improves manipulation technique detection, 2) boosts confidence, 3) improves ability to discern trustworthy & untrustworthy content, and 4) improves the quality of people's sharing decisions 3/12 https:…
This campaign is targeting an audience as large as the entire population of these three countries. And Jigsaw, Google's research arm, will be studying the campaign's effects along the way. https://twitter.com/...
The goal, according to @yasmind of Jigsaw, which is the Google group leading this work, is to move beyond fact-checking specific lies and to move toward building broader resilience against the tropes and tactics used to push a whole bunch of lies. https://www.protocol.com/...
The paper found that when people watched these inoculation videos - both in a lab and in the wild on YouTube - they were more likely to correctly spot manipulative content later on and resist sharing it. https://www.protocol.com/...