The tragic death of 16-year-old Vincent has thrust British online safety campaigners into the spotlight, demanding platform accountability after his parents revealed that algorithms promoted content reinforcing a message that he was ‘never good enough’. As a scientist who tracks data streams as assiduously as social media feeds, I find the parallels between the two domains disquieting. Both rely on feedback loops. In climate systems, positive feedback accelerates warming: melting ice reduces albedo, absorbing more heat. In social platforms, engagement-driven algorithms amplify negative content, creating a psychological feedback loop for vulnerable users. The Vincent case exemplifies this: his parents have stated that the algorithm served increasingly harmful material, validating a distorted self-perception until it became a narrative of despair.
Data from Ofcom indicates that 65% of UK teens have encountered harmful content online. Yet platforms argue that personalising content is their raison d‘être. The analogy with energy transitions is stark: we cannot solve a problem using the same thinking that created it. Relying on platforms to self-regulate is akin to expecting fossil fuel companies to lead the decarbonization effort. Self-interest invariably trumps societal good. The Online Safety Bill, passed into law last year, imposes a duty of care on platforms, but enforcement remains slow. Where are the fines? Where are the structural changes to algorithmic design?
Campaigners are now calling for independent audits of recommendation systems, transparency reports, and a ‘safety by design’ mandate. This mirrors the push for carbon budgets and net-zero targets: we need measurable, enforceable metrics. The Vincent family wants to see platforms block harmful content before it reaches children. But this is technically challenging. Filtering algorithms often struggle with nuance, and over-blocking risks censorship. However, the alternative is unacceptable. The biosphere cannot afford indefinite emissions; similarly, a generation cannot afford indefinite exposure to algorithmic harm.
The urgency in this case is palpable. Each day without action, thousands more children are shaped by algorithms optimised for engagement, not wellbeing. We have the technological solutions: better AI, user-controlled settings, mandatory risk assessments. What we lack is the political and corporate will. The Vincent tragedy should be the equivalent of a climate tipping point. If we do not act now, the psychological climate for young people will deteriorate beyond repair. The data is clear. The science is settled. Now, we need accountability.
