“Sam Altman has acknowledged wanting to train on everyone's personal documents (Word files, email etc)”
Our reading, the sentence that was checked:
Sam Altman has acknowledged that he wants to train AI on everyone's personal documents, such as Word files and email.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, OpenAI: On a Path to Becoming The World’s Most Frightening Surveillance Company?, paragraph 2.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-openai-on-a-path-to-becoming-the-c9ddff-a1
We recorded 6 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2024-08-07. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
From the speaker
- Sam Altman has acknowledged
- wanting to train
- everyone's personal documents
- Word files, email etc
Our interpretation
- The claim attributes a public statement to Altman made by August 7, 2024.
- I read 'everyone's' as users generally, without adding a claim that participation would be compulsory.
- Training on personal documents is distinct from accessing documents through retrieval or a context window.
- Discussing a possible future does not by itself establish an acknowledged intention to implement it.
Counts as supported if: A dated statement by Altman explicitly expresses a desire or intention for AI to learn through training on users' personal documents and email.
Counts as contradicted if: An authoritative correction or the complete attributed statement establishes that Marcus materially reversed or misrepresented what Altman said. Failure to locate a matching statement alone is insufficient.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Unresolved
The record was checked, and the evidence available by the cutoff does not settle it.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
Altman publicly described highly personalized assistants and discussed voluntary training on a person's life. Those remarks give Marcus's paraphrase a factual basis. They do not clearly establish the stronger attribution that Altman acknowledged wanting to train on everyone's personal documents. Treating access, context, and training as interchangeable would determine the result. The located evidence also does not justify calling the attribution false.
Limitations
- The search cannot establish that no other qualifying statement exists.
- Public remarks support a broader personalization reading more clearly than the specific training claim.
Evidence (3 items)
Context only: MIT News,
privacy versus utility
Why it matters here: The surrounding paragraph reports Altman discussing advantages if someone permits training on their entire life, while saying the answers remain uncertain. This establishes discussion of voluntary personal-data training, but does not explicitly establish the full intention Marcus attributes.
- Publisher
- MIT News
- Kind of source
- News report
- Describes events of
- Public from
Context only: Lex Fridman Podcast,
user choice
Why it matters here: Altman describes putting his email and conversations into a large context window, then emphasizes control over what an assistant remembers. This supports interest in personalized access, while leaving the claim about training unestablished.
- Publisher
- Lex Fridman Podcast
- Kind of source
- News report
- Describes events of
- Public from
Context only: MIT Technology Review,
off helping you
Why it matters here: The interview describes Altman's desired assistant as knowing his life, email, and conversations. It does not specify training on everyone's documents as the mechanism.
- Publisher
- MIT Technology Review
- Kind of source
- News report
- Describes events of
- Public from
What would settle it
The original recording or transcript Marcus had in mind, with language establishing Altman's intention and whether he meant training or access.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, OpenAI: On a Path to Becoming The World’s Most Frightening Surveillance Company?
Sam Altman has acknowledged wanting to train on everyone's personal documents (Word files, email etc)
- Where
- paragraph 2
- Published
- 2024-08-07
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-openai-on-a-path-to-becoming-the-c9ddff-a1: Current. Outcome: Unresolved. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.
About assessment marcus-openai-on-a-path-to-becoming-the-c9ddff-a1.