ClaimsEliezer Yudkowsky

“Or if you're tracking the destination of the hundreds of millions of dollars AI companies are pouring into lobbying and PACs, to try to primary any politicians who favor regulating AI.”

Eliezer Yudkowsky, author on X 1 April 2026 Post by @ESYudkowsky on X, 2026-04-01 (X)

Factual AI policy and governance

AI companies are pouring hundreds of millions of dollars into lobbying and PACs to try to primary politicians who favor regulating AI.

Unresolvable The public record could not settle it.

Factual: Statements about the present or the past, checked against the record: did they hold up?

The second judge would not resolve this claim; the verdict shown is the first judge's.

What would count as true or false

Judge as of 2026-04-01. True: credible reporting or FEC/lobbying disclosures show that AI companies, their executives or affiliated super PACs (e.g. Leading the Future) committed or spent a combined total in the hundreds of millions of dollars (at least about $200M) on lobbying and PACs, and that part of this money was aimed at defeating politicians, including in primaries, who favour regulating AI. Mostly true: the total reaches the hundreds of millions only when pledged or committed funds are counted alongside actual spending, or the money comes partly from AI investors and executives rather than the companies themselves, while the anti-regulation electoral targeting is documented. Mixed: the money is documented at that scale but the targeting of pro-regulation politicians is not, or the targeting is documented but the sums are well below $100M. Mostly false: the spending is in the tens of millions, or is not aimed at challenging pro-regulation politicians. False: there is no significant AI-industry PAC or lobbying effort of that kind.

Set before the research began.

Evidence (4)

  1. E1 Both sides claim the lead in AI’s high-stakes midterm race Transformer · 10 March 2026

    “Organizations focused on AI have, as of February, raised more than $175 million to spend on midterm Congressional candidates. Only $9.36 million of that has been spent.”
  2. E2 Both sides claim the lead in AI’s high-stakes midterm race Transformer · 10 March 2026

    “That’s despite $1.95m in Bores attack ads from the Leading the Future network, a “pro-innovation” group of super PACs funded by OpenAI president Greg Brockman and venture capital firm Andreessen Horowitz, among others.”
  3. E3 The AI industry is all in for the 2026 midterms with government regulations looming ABC News · 2 April 2026

    “As of March 16, a super PAC tied to Leading the Future had already spent more than $2.2 million opposing Bores, FEC filings show.”
  4. E4 The AI industry is all in for the 2026 midterms with government regulations looming ABC News · 2 April 2026

    “An AI-related political group, Innovation Council Action, tied to two of President Donald Trump's advisors, announced on Sunday that it would spend at least $100 million, The New York Times reported.”

Reasoning

The evidence shows AI-focused groups raising more than $175M (E1), another AI-related group announcing at least $100M (E4), and the industry-funded Leading the Future network spending about $2M against Alex Bores (E2, E3). None of the evidence shows that Bores or any other targeted politician favours regulating AI, so it does not establish the targeting the criteria require.

Cite this claim

Eliezer Yudkowsky, “AI companies are pouring hundreds of millions of dollars into lobbying and PACs to try to primary politicians who favor regulating AI.” (said 1 April 2026). Verdict: unresolvable. Reliability API, claim c_8f9189b1448f, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_8f9189b1448f

This claim is in dataset release v1.1.2, with the same verdict.

Reply to this claim

For Eliezer Yudkowsky, or someone speaking for them: a correction or counter-evidence. Replies are read before anything is shown; an approved reply shows your name, role, date, statement and link. Your email address is never shown and is deleted 30 days after the decision (privacy).

More factual claims by Eliezer Yudkowsky

  1. Mixed Factual 20 April 2026
    “I was observing that obvious-seeming hypothesis myself, sometime around the time of GPT 3.5.”

    Yudkowsky himself observed the persona-selection-like hypothesis around the time of GPT 3.5.

  2. Mostly true Factual 20 April 2026
    “The test-passing preference stands on its own once formed.”

    Once formed through RL training, an LLM's test-passing preference stands on its own and is not overridden by its inference of what the user wants.

  3. Mostly true Factual 20 April 2026
    “LLMs surely have neural circuits they learned to model text streams from fictional and nonfictional personas that are not lying, deceiving, cheating.”

    LLMs surely have neural circuits they learned to model text streams from fictional and nonfictional personas that are not lying, deceiving…

  4. Mostly true Factual 20 April 2026
    “The predict-as-if-honest circuitry is contextual, invoked to predict some streams of text but not others.”

    LLMs' predict-as-if-honest circuitry is contextual, invoked for predicting some text streams but not others.

  5. Mostly true Factual 20 April 2026
    “Presumably, somewhere along the way LLMs learn a dispreference for tests that don't pass, independently of how that gets prevented or whether the new plan does what the user really wanted.”

    Presumably, at some point during RL training on code-passing tasks, LLMs learn a dispreference for failing tests, independent of how that is avoided or whether the user's real goal is met.

All of them

Verdicts are made automatically; how verdicts are made.