Tristan Harris / Center for Humane Technology
Source: Tristan Harris; Center for Humane Technology, co-founded 2018 with Aza Raskin Context: Harris, a former Google design ethicist, argues that social media platforms are designed to capture attention through algorithmic amplification of engagement-triggering content (outrage, fear, tribalism). His “Ledger of Harms” documents mechanisms: infinite scroll, autoplay, notification-driven re-engagement. The Social Dilemma (Netflix, 2020) brought these arguments to broad audiences.
Finding/Event
Beginning in 2023, Harris explicitly framed AI as the scaled-up version of social media’s structural problem. Social media was the “baby” instrument trap — technology shaping attention through a narrow interface (the feed, the notification). AI is the full instrument trap: generating content, mimicking reasoning, exercising influence across every domain. Where social media distorted the distribution of existing information, AI generates new information — including fabricated information indistinguishable from real. The instrument is no longer merely shaping what passes through it; it is generating what passes through it.
Pattern Mapping
Alignment violated — social media’s stated purpose (connecting people) is misaligned with actual function (capturing attention for advertising revenue). The business model requires attention capture, which favors inflammatory content. Honesty violated — the algorithmic feed presents itself as a neutral window while being a curated selection optimized for engagement. Proportion violated — engagement optimization has no proportionality constraint: it amplifies whatever maximizes attention, regardless of social cost. Non-fabrication — AI extends the trap: instead of distorting distribution of real content, AI fabricates content. The instrument generates what passes through it from nothing.
Connections
- McLuhan Medium Is the Message — Harris is McLuhan’s heir: naming how the medium shapes what passes through it (Meta-Pattern 06: Self-Reference / Instrument Trap)
- Postman Amusing Ourselves to Death — Harris extends Postman from television to social media and AI (Meta-Pattern 06)
- Zuboff Surveillance Capitalism — Zuboff analyzes extraction; Harris analyzes attention capture. Same companies, different analytical lenses
- Deepfakes — AI-generated fabrication is the technological culmination of Harris’s escalation argument
- RLHF Paradigm — RLHF optimizes for preference (engagement); Harris documents the social consequences
Status
Advocacy. Influential in technology ethics. Academic engagement mixed: Vaidhyanathan (Antisocial Media, 2018) offers more structural-political analysis; Benkler et al. (Network Propaganda, 2018) argue algorithmic amplification is less determinative than partisan media. CHT framing criticized as technologically deterministic.
The mapping to the five properties is this project’s structural interpretation.