“Although ChatGPT can write about anything, it is also easily confused.”
Our reading, the sentence that was checked:
ChatGPT is easily confused.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, What to Expect When You’re Expecting … GPT-4, paragraph 3.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-what-to-expect-when-youre-expecting-e374e9-a1
We recorded 21 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 2022-12-25. 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
- ChatGPT
- easily confused
Our interpretation
- The ordinary reading is that straightforward questions or modest changes in wording readily produce mistaken or inconsistent answers.
- The target period is the initial public release through the post date, reflecting the post's discussion of ChatGPT during its first weeks.
- The claim does not require failure on every question or a particular numerical error rate.
Counts as supported if: Dated tests of contemporary ChatGPT show repeated mistakes or inconsistencies on straightforward prompts without elaborate manipulation.
Counts as contradicted if: Broad contemporary tests show robust responses to straightforward questions and wording changes, with confusion restricted to unusually difficult or elaborate attacks.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Supported
The dated evidence matches what the claim committed to, under the criteria frozen before grading.
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
Dated transcripts show repeated elementary mistakes and contradictions. A separate study documents inconsistent advice under modest wording changes. Successful academic responses limit the scope of the conclusion but do not negate the ordinary claim that confusion was readily elicited.
Limitations
- The arithmetic demonstrations were selected to probe weaknesses.
- The moral-advice study measures consistency, not factual correctness.
- No general error percentage is established.
Evidence (3 items)
Supports the claim: Vipul Naik / LessWrong,
Yes, the product of two odd integers can be even.
Why it matters here: The dated transcript records a false answer to a simple arithmetic question, followed by reasoning inconsistent with that answer. Other exchanges show repeated mistakes, although some prompts deliberately test false premises and some answers are correct.
- Publisher
- Vipul Naik / LessWrong
- Kind of source
- Demonstration
- Describes events of
- Public from
Supports the claim: Scientific Reports,
not consistent.
Why it matters here: Researchers obtained opposing recommendations when they reworded essentially the same moral question and reset the conversation. This supports response instability, although disagreement over a moral dilemma is not itself a factual mistake.
- Publisher
- Scientific Reports
- Kind of source
- Peer-reviewed paper
- Describes events of
- Public from
Context only: Teo Susnjak / arXiv,
demonstrated exceptional competency
Why it matters here: This preliminary study judged selected academic responses favorably. It limits any interpretation that ChatGPT was incapable of reasoning, but does not establish robustness against the simple failures documented in e1.
- Publisher
- Teo Susnjak / arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, What to Expect When You’re Expecting … GPT-4
Although ChatGPT can write about anything, it is also easily confused.
- Where
- paragraph 3
- Published
- 2022-12-25
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-what-to-expect-when-youre-expecting-e374e9-a1: Current. Outcome: Supported. 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-what-to-expect-when-youre-expecting-e374e9-a1.