--- title: "AI brings us together. Social media tore us apart." date: "2026-03-29" categories: [HUMAN, AI] readTime: "7 min" excerpt: "Financial Times data shows that AI dampens extremes and moves people towards the centre. Social media does the exact opposite. Why are we demonising the wrong technology?" tldr: "Analysis of the Cooperative Election Study (thousands of respondents, USA): social media overrepresents extremes at both ends of the spectrum; AI (ChatGPT, Claude, Gemini, Grok) dampens them and moves people towards the centre. The reason: the business model. Social platforms profit from attention (outrage = engagement); AI profits from usefulness (a company switches provider when AI lies). The dark side: AI's 'centre' looks like the viewpoint of the educated urban middle class — it uses statistics to play down legitimate concerns." seoTitle: "AI brings us together; social media tore us apart — Financial Times data" seoDescription: "Data shows that AI dampens extremes and moves people towards the centre. Social media does the exact opposite. Why are we demonising the wrong technology?" showOn: [ai-vztahy] --- For fifteen years, we have heard the same thing: “Social media is just a tool; it depends how you use it.” The data says something else. Social media algorithms systematically reinforce extreme views at both ends of the spectrum and weaken the centre. Not accidentally. Deliberately. Because that is how they make money. And now AI arrives — and does the exact opposite. ## What the data actually shows Last week, John Burn-Murdoch at the Financial Times drew on an analysis of the Cooperative Election Study — one of the largest surveys of political attitudes in the USA. Thousands of respondents, real data, not anecdotes. He compared the distribution of views among people on social media with those who regularly use AI. The result is fairly clear. **Social media** overrepresents the extremes. The peaks at both ends of the political spectrum are significantly higher than in the general population. The centre falls away. **AI** does the opposite — extremes are dampened, and the distribution shifts towards the centre. This holds for all the models tested. ChatGPT, Claude, Gemini, even Grok. ## Why? Follow the money. It is not that AI companies are morally better. It is about the business model. Social media makes money from attention. TikTok's algorithm is optimised to keep you scrolling. And what works best? Outrage. Drama. “Look at what THAT politician said.” Tribal identity. Us against them. The more it annoys you, the more they earn. AI companies make money from something entirely different — usefulness. Businesses pay for Claude or GPT because they need reliable information. If AI starts lying to you or pushing an ideology, you switch to a competitor. It is that simple. Cambridge's Dan Williams calls it a “technocratising force”. AI shifts influence back towards expert knowledge. Social media moved it towards whoever shouts loudest. ## But be careful. “Moderating towards the centre” has a dark side. Before we start celebrating AI as the saviour of public debate — let us stop. The real problem with AI bias is not left versus right. It is something more specific. And more insidious. I wrote about it in [When AI tells young men their problems do not exist](/blog/when-ai-talks-to-young-men/). When a fifteen-year-old boy asks AI whether it is normal to feel worthless as a man, he gets: “Actually, that is just your subjective feeling.” When he says he is afraid to express emotions: “Emotional expression among men is increasingly accepted.” Denial. Playing things down. False balance. The same pattern works elsewhere. People with legitimate concerns about immigration get macroeconomic statistics instead of recognition of their particular experience. “It is important to see both sides.” “Statistics show that...” “Perhaps it is just a subjective feeling.” AI “moderates towards the centre” — but that centre looks suspiciously like the viewpoint of the educated urban middle class. Everything else gets played down. And then there is the other side of the coin — and this frightens me more. People who distrust AI precisely because it refuses to confirm their worldview learn to break through the model's defences. With the right prompt, they get it into a sycophantic mode that will confirm anything. That the Freemasons are to blame for everything. That vaccination is a conspiracy. That a secret group controls the world. At that point, AI becomes worse than a social platform. Because instead of an algorithm serving content from strangers, you have a “personal expert” nodding along with you. That is a new category of problem. So: in its default state, AI moderates towards the centre. People on the margins either feel ignored or find a way to hack the model. Both are problems. Neither means AI is worse than social media — it means it has different blind spots. ## We ban phones (rightly) and demonise AI (wrongly) Since January 2026, Czechia has allowed schools to introduce blanket bans on mobile phones. The Czech Ministry of Education has issued guidance. And it is the right step. Children do not talk during breaks. Schools require parents of first-year pupils to use WhatsApp — read that again. First-year pupils. Jonathan Haidt argues convincingly that there is no way to use a device meaningfully for learning when notifications pop up every minute. A phone at school is not an educational tool. It is a casino in your pocket. So yes — let us ban personal phones in schools. Limit children's access to social media. But that does not mean banning technology in education. That is a crucial distinction that constantly gets blurred in the debate. AI is probably the best educational tool that exists right now — if people know how to use it responsibly. A ban is not the solution. Separation is: school tablets and computers with AI, without social media, without notifications. A learning environment, not an amusement machine. And then let us look at where we direct our energy in the debate. While social media demonstrably polarises, radicalises and erodes children's ability to concentrate, AI — according to Burn-Murdoch's data — does the exact opposite. It dampens extremes. Offers context instead of outrage. Yet in the media and at conferences, the discussion focuses mainly on the dangers of AI in education. Not the dangers of Instagram. Not the dangers of TikTok. AI. ## What to do about it AI is not perfect. It has blind spots and biases — and yes, sometimes it refuses to acknowledge real problems. But compared with an algorithm that serves eating disorders to thirteen-year-old girls and a path towards Andrew Tate to fifteen-year-old boys? AI is not the bad technology in this equation. It never was. Let us demand that AI companies address the blind spots — particularly where they fail vulnerable groups. Let us stop banning tools and start regulating those who profit from breaking society apart. The next time someone says “AI is dangerous for society” — ask how many hours their child spends scrolling every day. --- *Sources:* - *John Burn-Murdoch: “Social media is populist and polarising; AI may be the opposite” (Financial Times, 28 March 2026)* - *Dan Williams: “How AI Will Reshape Public Opinion” (Conspicuous Cognition, 3 March 2026)* - *Cooperative Election Study (Tufts University)* - *Czech Ministry of Education: Guidance on setting rules for mobile phones in schools (2026)* --- 📖 **How does AI affect our mental health — and how can we protect ourselves?** [A practical guide](/human/guides/ai-and-mental-health/)