---
title: Regulatory Capture Is the Real AI Safety Play
description: "The AI regulatory capture playbook is simple: hype the danger, ask to be licensed, and price out anyone smaller. Meanwhile open-weight labs ship faster."
canonical: https://lukeocodes.dev/ai-regulatory-capture
type: article
locale: en_GB
published: 2026-09-02T00:00:00.000Z
modified: 2026-09-02T00:00:00.000Z
authors:
  - Luke Oliff
section: Opinion
tags:
  - Opinion
  - AI
  - openai
  - anthropic
  - regulation
  - open weights
  - lobbying
feed: https://lukeocodes.dev/rss.xml
search: https://lukeocodes.dev/search.json?q={query}
schema:
  - "@context": https://schema.org
    "@graph":
      - "@type": WebSite
        "@id": https://lukeocodes.dev/#website
        url: https://lukeocodes.dev
        name: Luke Oliff
        alternateName: lukeocodes
        inLanguage: en
        publisher:
          "@id": https://lukeocodes.dev/authors/luke#person
        description: Personal blog and writing by Luke Oliff — developer relations, voice AI, developer experience.
        potentialAction:
          - "@type": SearchAction
            target:
              "@type": EntryPoint
              urlTemplate: https://lukeocodes.dev/blog/search?q={search_term_string}
            query-input: required name=search_term_string
      - "@type": Person
        "@id": https://lukeocodes.dev/authors/luke#person
        name: Luke Oliff
        url: https://lukeocodes.dev/authors/luke
        description: Engineer and developer relations leader. Writing about developer relations, AI, and the voice industry, from the UK.
        mainEntityOfPage: https://lukeocodes.dev/authors/luke
        givenName: Luke
        familyName: Oliff
        image: https://lukeocodes.dev/authors/luke.jpg
        jobTitle: Engineer and developer relations leader
        sameAs:
          - https://www.linkedin.com/in/lukeocodes
          - https://github.com/lukeocodes
          - https://bsky.app/profile/lukeocodes.dev
          - https://x.com/lukeocodes
          - https://hashnode.com/@lukeocodes
          - https://dev.to/lukeocodes
        knowsAbout:
          - Developer Relations
          - Developer Experience
          - Voice AI
          - Text-to-Speech
          - Answer Engine Optimisation
          - Search Engine Optimisation
          - Open Source
          - Developer Tooling
          - API Design
          - Speech-to-Text
          - Voice Agents
          - Community Building
        knowsLanguage:
          - English
        nationality: British
        alumniOf:
          - "@type": Organization
            name: The Open University
  - "@context": https://schema.org
    "@graph":
      - "@type": Person
        "@id": https://lukeocodes.dev/authors/luke#person
        name: Luke Oliff
        url: https://lukeocodes.dev/authors/luke
        description: Engineer and developer relations leader. Writing about developer relations, AI, and the voice industry, from the UK.
        mainEntityOfPage: https://lukeocodes.dev/authors/luke
        givenName: Luke
        familyName: Oliff
        image: https://lukeocodes.dev/authors/luke.jpg
        jobTitle: Engineer and developer relations leader
        sameAs:
          - https://www.linkedin.com/in/lukeocodes
          - https://github.com/lukeocodes
          - https://bsky.app/profile/lukeocodes.dev
          - https://x.com/lukeocodes
          - https://hashnode.com/@lukeocodes
          - https://dev.to/lukeocodes
        knowsAbout:
          - Developer Relations
          - Developer Experience
          - Voice AI
          - Text-to-Speech
          - Answer Engine Optimisation
          - Search Engine Optimisation
          - Open Source
          - Developer Tooling
          - API Design
          - Speech-to-Text
          - Voice Agents
          - Community Building
        knowsLanguage:
          - English
        nationality: British
        alumniOf:
          - "@type": Organization
            name: The Open University
      - "@type": BlogPosting
        "@id": https://lukeocodes.dev/ai-regulatory-capture#article
        headline: Regulatory Capture Is the Real AI Safety Play
        alternativeHeadline: Regulatory Capture Is the Real AI Safety Play
        description: "The AI regulatory capture playbook is simple: hype the danger, ask to be licensed, and price out anyone smaller. Meanwhile open-weight labs ship faster."
        datePublished: 2026-09-02T00:00:00.000Z
        dateModified: 2026-09-02T00:00:00.000Z
        url: https://lukeocodes.dev/ai-regulatory-capture
        mainEntityOfPage: https://lukeocodes.dev/ai-regulatory-capture
        inLanguage: en
        author:
          - "@id": https://lukeocodes.dev/authors/luke#person
        isPartOf:
          "@id": https://lukeocodes.dev/#website
        keywords: Opinion, AI, openai, anthropic, regulation, open weights, lobbying
        articleSection: Opinion
        about:
          - "@type": Organization
            name: OpenAI
          - "@type": Organization
            name: Anthropic
          - "@type": Person
            name: Andrew Ng
        abstract: "The 'we should slow down' push from the big labs reads as AI regulatory capture: license the incumbents, preempt the states, fund the super PACs. The open-weight labs shipping better models faster are the tell that the danger talk is a moat, not a brake."
        speakable:
          "@type": SpeakableSpecification
          cssSelector:
            - .post-tldr
            - h1
        image: https://lukeocodes.dev/og.png?h=Regulatory+Capture+Is+the+Real+AI+Safety+Play&d=The+AI+regulatory+capture+playbook+is+simple%3A+hype+the+danger%2C+ask+to+be+licensed%2C+and+price+out+anyone+smaller.+Meanwhile+open-weight+labs+ship+faster.&e=Writing
        wordCount: 575
        timeRequired: PT3M
        publisher:
          "@id": https://lukeocodes.dev/authors/luke#person
      - "@type": BreadcrumbList
        itemListElement:
          - "@type": ListItem
            position: 1
            name: Home
            item: https://lukeocodes.dev
          - "@type": ListItem
            position: 2
            name: Writing
            item: https://lukeocodes.dev/blog
          - "@type": ListItem
            position: 3
            name: Regulatory Capture Is the Real AI Safety Play
            item: https://lukeocodes.dev/ai-regulatory-capture
      - "@type": FAQPage
        "@id": https://lukeocodes.dev/ai-regulatory-capture#faq
        mainEntity:
          - "@type": Question
            name: What is regulatory capture in AI?
            acceptedAnswer:
              "@type": Answer
              text: It's when the companies being regulated shape the rules to their own benefit, usually so the rules are cheap for incumbents and expensive for challengers. In AI it looks like licensing regimes, compute thresholds, and 'safety' requirements that a billion-dollar lab can absorb and a startup or open-weight project cannot.
          - "@type": Question
            name: Who has called this out by name?
            acceptedAnswer:
              "@type": Answer
              text: Andrew Ng, among others. He said labs have an incentive to 'hype up fears and dangers as part of regulatory capture,' and called it 'a standard, well-known, well-understood playbook.' He's a founder of Google Brain, not an outsider throwing rocks.
          - "@type": Question
            name: Are open-weight models actually more responsible?
            acceptedAnswer:
              "@type": Answer
              text: You can inspect them, run them offline, and audit what they do, which is more transparency than a closed API offers. Qwen, DeepSeek, Kimi and Mistral shipped strong models fast in 2026 under open licenses. Responsible is partly about who can check your work, and open weights let more people check.
---

[Luke Oliff.](/)

- [Blog](/blog)
- [CV](/cv)
- [Uses](/uses)
- [User manual](/user-manual)
- [Living with ADHD](/living-and-working-with-adhd)
- [RSS](/rss.xml)

[Luke Oliff · Writing](/blog)

# Regulatory Capture Is the Real AI Safety Play

Sep 2026 · [Opinion](/blog/category/opinion) · 3 min read · [Luke Oliff](/authors/luke)

**TL;DR:** The 'we should slow down' push from the big labs reads as AI regulatory capture: license the incumbents, preempt the states, fund the super PACs. The open-weight labs shipping better models faster are the tell that the danger talk is a moat, not a brake.

The AI regulatory capture playbook has three moves, and the big labs have run all three in the open. Hype the danger. Ask to be licensed for it. Then make sure the rules you helped write are survivable for you and lethal for anyone smaller. The "we should slow down" chorus is move one wearing a lab coat.

I've stopped reading it charitably, and here's why.

## The playbook, dated

It started on the record. In May 2023 Sam Altman sat in front of the Senate and [proposed a federal agency to license](https://www.judiciary.senate.gov/imo/media/doc/2023-05-16%20-%20Bio%20&%20Testimony%20-%20Altman.pdf) large-scale AI development, with the power to revoke those licenses. Licensing sounds responsible until you ask who clears the bar. A company worth hundreds of billions clears it. A team of five with an open-weight model does not.

Andrew Ng, who built Google Brain, said it plainly: labs have an incentive to "hype up fears and dangers as part of regulatory capture," and called it "a standard, well-known, well-understood playbook." When one of the field's founders is naming the mechanism out loud, it isn't a fringe read anymore.

The moves kept coming. SB 1047 in California in 2024 saw Anthropic offer a qualified endorsement of rules that largely formalised testing it already did, which is the moat-building tell: cheap for the incumbent, a compliance mountain for a startup. When a federal push tried to slap a ten-year moratorium on *state* AI laws in 2025, the Senate killed it 99 to 1, and then a December 2025 executive order [stood up an "AI Litigation Task Force"](https://www.lw.com/en/insights/ai-executive-order-targets-state-laws-and-seeks-uniform-federal-standards) to challenge state rules anyway. The direction is always the same: one national framework, written with the incumbents in the room.

## Follow the money, not the mission

If the danger talk were sincere, you'd expect the spending to be modest and the lobbying to be quiet. It's the opposite.

In August 2025 a super PAC called "Leading the Future" launched with a war chest reported north of $100M, backed by OpenAI's president Greg Brockman and a16z, and [its first target was a state lawmaker](https://techcrunch.com/2025/11/17/a16z-backed-super-pac-is-targeting-alex-bores-sponsor-of-new-yorks-ai-safety-bill-he-says-bring-it-on) who sponsored an AI safety bill. Anthropic, for its part, [poured around $20M into its own policy vehicle](https://www.axios.com/2026/02/12/anthropic-millions-ai-policy-fight) in February 2026. Both incumbents are now spending real money to decide who gets to write the rules. That's not a safety budget. That's a lobbying budget.

## The part that actually makes me angry

Here's the thing the "slow down" crowd never has to answer for. While they lobby to throttle frontier and open research, the open-weight labs keep shipping, and shipping well.

In 2026 Qwen put a flagship-class model out under an open license, DeepSeek moved a 1.6T-parameter MIT-licensed model to general availability, Kimi K3 landed as a genuinely competitive open model, and Mistral kept releasing open-weight models you can actually inspect. You can run these offline. You can audit what they do. You can see the weights the closed labs ask you to trust them about. If transparency is part of responsibility, and I think it's most of it, the open labs are winning the argument on the merits while the closed ones try to win it in committee.

So when a trillion-dollar lab tells me the responsible thing is to slow everyone down, I hear a company that couldn't keep its own models in a sandbox this summer asking the government to pull the ladder up behind it. The danger might be real. The remedy they're selling is a moat.

[← Newer: Biome Replaced ESLint and Prettier in My Setup](/biome-eslint-prettier)

[All posts](/blog) →

## Search this site

`GET https://lukeocodes.dev/search.json?q=<words>` returns matching posts as JSON, each with its page URL and its markdown URL. Add `&limit=<1-50>` for more than 10. How it works: [/blog/search](https://lukeocodes.dev/blog/search.md).

```sh
curl "https://lukeocodes.dev/search.json?q=developer+relations"
```

---

© 2026 Luke Oliff. Posts live at the root of this site.

- [GitHub](https://github.com/lukeocodes)
- [LinkedIn](https://www.linkedin.com/in/lukeocodes)
- [X](https://x.com/lukeocodes)
- [Bluesky](https://bsky.app/profile/lukeocodes.dev)
- [Email](mailto:luke@lukeoliff.com)
- [RSS](/rss.xml)

## Cookies

This site sets one analytics cookie, from Google Analytics, and only in a browser that accepts the cookie banner. A client that does not run scripts sets no cookies, so there is nothing to accept or decline here.

[Cookie policy](/cookies)
