“Eventually, we will be able to build machines that help us reason critically, overcome human bias, and objectively, the way that we have calculators that spit out just the facts, without confabulation.”
Our reading, the sentence that was checked:
Eventually, we will be able to build machines that help us reason critically, overcome human bias, and reason objectively, giving just the facts without confabulation, the way calculators do.
Assessment: Unresolved No deadline stated on 2026-10-29 Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, AI versus The Clumsy Human Mind, paragraph 19.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-ai-versus-the-clumsy-human-mind-43b7c4-a1
We recorded 7 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-29. The source may have been edited after 2023-11-25. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Forecast
- Deadline, as the speaker wrote it
- “Eventually”
- Deadline state when assessed
- No deadline stated
- Deadline state on 2026-10-29
- No deadline stated (the release’s reference date)
From the speaker
- Eventually
- help us reason critically
- overcome human bias
- just the facts, without confabulation
Our interpretation
- The plain reading is that machines will combine useful critical-reasoning assistance, reduced human bias, and dependable factual answers across varied ordinary tasks.
- The calculator comparison implies dependable factual reliability. It does not require answering unknowable questions or prohibit admitting uncertainty.
- Success on one narrow task does not establish the combined capability. No deadline or numerical reliability threshold was supplied.
Counts as supported if: A built system demonstrably combines improved human critical reasoning, reduced bias, and factual reliability without confabulation across varied tasks, confirmed by independent evaluations.
Counts as contradicted if: A decisive result establishes that the combined capability cannot be built. Failures of existing systems cannot by themselves contradict an undated eventual forecast.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Did the forecast come true?
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 deadline stated
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
There is evidence of progress toward critical-thinking assistance, while factual unreliability persists in evaluated language models. Neither establishes the full combined capability or rules out its eventual construction. Marcus supplied no deadline, and no decisive event meeting the early-resolution rule was found.
Limitations
- Small, task-specific studies cannot establish broad capability.
- Evidence about current language models does not cover every possible machine architecture.
- The breadth and reliability requirements involve evaluator interpretation.
Evidence (2 items)
Context only: arXiv,
We evaluated ArguMentor via a within-subjects experiment (N=24).
Why it matters here: The study reports improved argument generation and critical-thinking performance when participants used ArguMentor. This supports a component of the forecast, but the small study does not establish broad bias reduction or confabulation-free answers.
- Publisher
- arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: Nature,
Large language models sometimes produce confident, plausible falsehoods (‘hallucinations’), limiting their reliability1,2.
Why it matters here: The paper documents continuing hallucinations and proposes changes to evaluation incentives. It does not establish that future machines cannot meet Marcus's forecast.
- Publisher
- Nature
- Kind of source
- Peer-reviewed paper
- Describes events of
- Public from
What would settle it
Independent, replicated evaluations of a built system showing improved critical reasoning, reduced human bias, and dependable factual output together across varied tasks.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, AI versus The Clumsy Human Mind
Eventually, we will be able to build machines that help us reason critically, overcome human bias, and objectively, the way that we have calculators that spit out just the facts, without confabulation.
- Where
- paragraph 19
- Published
- 2023-11-25
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-ai-versus-the-clumsy-human-mind-43b7c4-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-ai-versus-the-clumsy-human-mind-43b7c4-a1.