ClaimsEliezer Yudkowsky

“The Dartmouth Proposal makes no mention of building nice/good/benevolent AI.”

Eliezer Yudkowsky, MIRI research fellow, essay author 2008 Artificial Intelligence as a Positive and Negative Factor in Global Risk (MIRI)

Factual AI policy and governance

The Dartmouth Proposal on AI makes no mention of building nice, good or benevolent 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.

Said for an organisation they lead or work for; it counts on both records.

What would count as true or false

Judge as of 2008, against the 1955 text of 'A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence'. True: the full text contains no mention of making AI nice, good, benevolent, safe or beneficial, or of its ethical or value alignment. Mostly true: the text has at most a passing phrase that could arguably touch on beneficial use but does not address building benevolent AI. Mixed: the text has some ambiguous references to good or beneficial AI. Mostly false: the text explicitly mentions the goal of benevolent or beneficial AI, though only briefly. False: the proposal clearly discusses building nice, good or benevolent AI.

Set before the research began.

Evidence (7)

  1. E1 A PROPOSAL FOR THE DARTMOUTH SUMMER RESEARCH PROJECT ON ARTIFICIAL INTELLIGENCE Stanford University (John McCarthy's history site, www-formal.stanford.edu) · 31 August 1955

    “For the present purpose the artificial intelligence problem is taken to be that of making a machine behave in ways that would be called intelligent if a human were so behaving.”

  2. E2 A PROPOSAL FOR THE DARTMOUTH SUMMER RESEARCH PROJECT ON ARTIFICIAL INTELLIGENCE Stanford University (John McCarthy's history site, www-formal.stanford.edu) · 31 August 1955

    “An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves.”

  3. E3 A PROPOSAL FOR THE DARTMOUTH SUMMER RESEARCH PROJECT ON ARTIFICIAL INTELLIGENCE Stanford University (www-formal.stanford.edu, John McCarthy's history pages) · 31 August 1955

    “An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves.”

  4. E4 A PROPOSAL FOR THE DARTMOUTH SUMMER RESEARCH PROJECT ON ARTIFICIAL INTELLIGENCE Stanford University (www-formal.stanford.edu, John McCarthy's history pages) · 31 August 1955

    “I am proposing here to start at the simple and when the environment is neither hostile (merely indifferent) nor complex, and to work up through a series of easy stages in the direction of these advanced activities.”

  5. E5 A PROPOSAL FOR THE DARTMOUTH SUMMER RESEARCH PROJECT ON ARTIFICIAL INTELLIGENCE Stanford University (www-formal.stanford.edu, John McCarthy's history pages) · 31 August 1955

    “5. Self-lmprovement”

  6. E6 Fermat's Library | A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence annotated/explained version. Fermat's Library · 1 January 2012

    “Probably a truly intelligent machine will carry out activities which may best be described as self-improvement.”

  7. E7 Editorial AI Magazine (American Association for Artificial Intelligence) · 2006

    “an excerpt from the Dartmouth Project proposal prepared by”

Reasoning

The evidence gives only short excerpts of the 1955 proposal (E1–E6). None of them mentions benevolent, good or safe AI, but they do not show that the full text contains no such mention, and that is what the criteria require for this verdict.

Cite this claim

Eliezer Yudkowsky, “The Dartmouth Proposal on AI makes no mention of building nice, good or benevolent AI.” (said 2008). Verdict: unresolvable. Reliability API, claim c_01bb9716f34b, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_01bb9716f34b

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.