Episode 67: Are Social Media Algorithms Designed to Radicalise Us?Facebook's own internal study found that sixty-four percent of users who joined extremist groups did so because the recommendation algorithm pushed those groups to them. Executives were shown proposals to fix it. They declined. Tonight we ask: is this a side effect of profit-seeking — or something more deliberately designed?This episode explores:* The leaked Carol Smith experiment — Facebook's fake account that received QAnon recommendations within two days* YouTube's documented rabbit hole — how recommendation systems pull users toward progressively more extreme content* A 2025 TikTok-like experiment showing that removing personalisation reduced harmful content by twenty-seven percent — but dropped usage by thirty-five percent* The European Union's Digital Services Act and why it requires platforms to assess radicalisation risk* The distinction between deliberate design for radicalisation and knowing tolerance of it for profit* What the academic research actually shows about the limits of the radicalisation narrativeFor believers, the leaked documents prove executives knew and chose engagement over safety — making the outcome as deliberate as if it were designed that way. For sceptics, profit-seeking and radicalisation are different things even when they overlap.What is certain: you are being shown what keeps you watching. What keeps you watching is not always what keeps you thinking clearly.Keywords: social media radicalisation conspiracy, Facebook algorithm extremism, YouTube rabbit hole radicalisation, are algorithms designed to divide us, social media manipulation truth, tech company radicalisation evidence, platform algorithm harm, TikTok radicalisation, Big Tech manipulation, engagement over safety

Ask this episode

“What did the council decide on item 42?” — answered with the speaker and the moment.

Transcript

Related