Google's new anti-extremist steps: targeted AI algo training, more content flaggers, no inflammatory video monetization, ads targeting potential ISIS recruits
Editor's Note: This post appeared as an op-ed in the Financial Times earlier today. — Terrorism is an attack on open societies …
Context & Ripple Effects
This op-ed lands three months after [[a:917463|the UK advertiser backlash forced Google to tighten ad placement controls and expand its hate speech definitions]], so the package reads as a second, more aggressive phase of that same brand-safety response — this time aimed at extremism itself rather than just where ads sit. The announcement bundles four levers at once: AI training on extremist content, a larger flagger team, demonetizing inflammatory videos, and redirecting advertising at potential ISIS recruits.
First-order effects
- Inflammatory videos on YouTube lose monetization immediately, cutting revenue to uploaders whose content falls short of explicit hate speech but feeds radicalization.
- Google's expanded flagger team and retrained algorithms begin actively surfacing counter-messaging ads to users identified as potential ISIS recruits.
Second-order effects
- Advertisers who pulled spend after the UK controversy gain a concrete demonstration that Google can police adjacent risk, pressuring rivals like Facebook to match both the demonetization rules and the counter-recruitment ad tactics.
- By running recruitment-targeted ads against ISIS's own audience, Google turns the ad auction from a passive placement system into an active intervention tool — a playbook other platforms must now decide whether to copy or reject.
Third-order effects
- The arc from this 2017 announcement runs through YouTube's later decision to broaden its extremist-content policy beyond violence and hate speech ([[a:924041]]) and culminates in Jigsaw and Tech Against Terrorism shipping Altitude as free terrorist-content detection for small platforms ([[a:846199]]) — moderation capability migrating from a single platform's internal rules to industry-shared infrastructure.
- If counter-extremism tooling keeps being distributed this way, smaller sites face a choice between adopting big-tech-supplied moderation stacks and bearing the compliance burden alone — a structural dependency regulators will increasingly scrutinize.
The trend: Platform counter-extremism work is evolving from reactive ad-policy patches into shared, AI-driven detection infrastructure that large platforms distribute across the wider web.