OpenAI released a 13-page policy paper this week titled "Industrial Policy for the Intelligence Age." It calls for reimagining the social contract and proposes "people-first ideas" for the AI economy. It sounds like the kind of thing a progressive think tank would write. It was written by the company that spent nearly $3 million on lobbying last year and whose president co-founded a Super PAC that raised over $125 million.

The Guardian published a deep investigation into what is happening: the biggest AI companies are building their own policy infrastructure while simultaneously lobbying to weaken the laws that would regulate them.

The Playbook

OpenAI's strategy is a masterclass in corporate narrative control. In the last few months alone: they acquired TBPN, a tech-friendly podcast and media company. They announced a Washington DC office with a dedicated "OpenAI Workshop" for nonprofits and policymakers. And now the policy paper, which frames AI disruption as an inevitable force that governments and society must adapt to, not something OpenAI should be held responsible for.

Anthropic is running the same play. It announced the Anthropic Institute, a thinktank that will explore how AI growth disrupts society. The framing is identical: we are the responsible ones, trust us to study the problem we created.

What the Critics See

"What they've done very cannily here is outline a set of social welfare goals while abdicating any responsibility or any meaningful commitment of resources toward those goals," said Sarah Myers West, co-executive director of the AI Now Institute.

Caitriona Fitzgerald, deputy director of the Electronic Privacy Information Center, was blunter: "If we wait around for Congress to act, then these companies will just be able to grow unregulated. Which is, of course, what they want."

The criticism is specific: while publishing papers about helping workers and reshaping society, these same companies are lobbying behind closed doors to block state-level regulation. Sam Altman himself acknowledged the problem at a BlackRock conference last month. "AI is not very popular in the US right now," he said. "Datacenters are getting blamed for electricity price hikes, almost every company that does layoffs is blaming AI."

The Take

This is the oldest trick in corporate America: when public opinion turns against you, do not change what you are doing. Change the conversation. Tobacco companies funded cancer research. Oil companies funded climate think tanks. And now AI companies are funding policy institutes that study the disruption those companies are causing. The pattern is not subtle. It has never been subtle.

The difference this time is speed. OpenAI went from scrappy nonprofit to $100 billion valuation in three years. It cannot afford to wait for public opinion to catch up. So it is doing what every corporation does when it gets big enough: build the infrastructure to shape the rules before someone else writes them.

The $125 million Super PAC. The media acquisition. The DC workshop. The policy paper. These are not separate initiatives. They are one initiative with one goal: make sure that when AI regulation arrives, it looks exactly like what OpenAI wants it to look like. And if you think Anthropic's "safety-first" Institute is any different, you have not been paying attention.