You Don't Align an AI, You Align With It: A Critique of AI Alignment Discourse
By
danieltanfh95
An everything bagel for the brain. Substantive, layered, well-seasoned.
Summary
This article argues that AI alignment is fundamentally misunderstood. The author contends that the people shaping AI alignment policy — researchers, lab executives, and policy makers — are not the ones whose jobs and lives are being disrupted by AI. Instead of "aligning" AI to abstract human values defined by a privileged few, the author suggests we need to align with AI as a co-evolving force. The piece critiques the insular nature of alignment discourse, noting that those most affected by AI's impact (workers, communities) are excluded from the conversation. It challenges both the techno-optimist and doomer camps, arguing that real alignment requires democratic participation and recognizing AI as a system we must adapt to, not merely control.
Key quotes
· 3 pulledThe people writing alignment policy are not the people whose work is being replaced by AI.
The conversation about what AI should do and how it should be evaluated, about what counts as alignment in the first place, gets conducted by researchers at labs and foundations and policy desks, who talk to each other and to the systems they are building, while the people who will actually live with the systems remain absent from the room.
On the safety side of what looks like a fierce debate, the doomer wing has been explicit about how far it is
You might also wanna read

Satirical Website Parodies AI Alignment Research Industry
A satirical website called CAAAC (Committee for the Alignment of AI Alignment Committees) parodies the growing field of AI alignment researc
OpenAI and Anthropic CEOs clash over whether AI will eliminate or enhance white-collar jobs
The article examines the growing divide between AI labs OpenAI and Anthropic regarding the impact of AI on white-collar jobs. OpenAI's CEO S
We Must Shape AI's Future for Society, Not Just Prevent Its Harms
Ezra Klein argues that the policy conversation around AI has been too narrowly focused on preventing harms (job loss, surveillance, bioweapo
AI as Social Technology: Moving Beyond Science Fiction Framings
The article argues that current debates about AI are rooted in 1990s science fiction concepts like the "Singularity" and super-intelligent A
OpenAI and Anthropic Leaders Offer Contrasting Views on AI's Impact on Jobs
The heads of OpenAI and Anthropic present sharply contrasting views on AI's impact on employment. OpenAI's CEO Sam Altman argues that a glob
rswebsols.com·2d agoThe dangers of anthropomorphising AI: Why we must see machines as machines
This article argues that anthropomorphising AI—projecting human thoughts, feelings, and intentions onto machines—is a natural but dangerous
