“ELIZA (built in 1965-1966) beat GPT 3.5.”
Our reading, the sentence that was checked:
In the cited Turing Test study, the 1960s ELIZA chatbot outperformed GPT-3.5.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Nope, the Turing Test has not been solved, paragraph 3.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-nope-the-turing-test-has-not-been-dd3f9e-a1
We recorded 4 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 2023-12-09. 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)
Conditions
- in the cited Turing Test study (arXiv 2310.20216) (qualifier, stated by the speaker)
From the speaker
- ELIZA
- beat GPT 3.5
- Full article at https://arxiv.org/abs/2310.20216
Our interpretation
- The comparison concerns the cited experiment, using the paper version available on December 9, 2023.
- Beat means a higher overall proportion of conversations in which interrogators judged the witness human.
- The target period is the 2023 experiment reported in the cited paper; exact collection dates are not specified in the passages reviewed.
- The historical parenthesis identifies ELIZA; it does not assert that the experiment ran the original executable.
Counts as supported if: The cited study reports a higher overall judged-human rate for ELIZA than for each tested GPT-3.5 configuration.
Counts as contradicted if: The cited study reports an overall judged-human rate for a tested GPT-3.5 configuration equal to or greater than ELIZA's.
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
The contemporary paper reports ELIZA at 27%, compared with 5% and 14% for the two GPT-3.5 prompts. Marcus accurately describes that experiment's overall ranking.
Limitations
- This establishes the reported sample ranking, not general superiority over GPT-3.5.
- Performance depended on prompts, and the study sampled configurations unevenly.
Evidence (1 item)
Supports the claim: arXiv; Cameron Jones and Benjamin Bergen,
GPT-3.5 performance ranged from 5% (Victor) to 14% (Juliet).
Why it matters here: The same results paragraph reports ELIZA at 27%, above both GPT-3.5 configurations. These are overall judged-human rates in the experiment Marcus cited.
- Publisher
- arXiv; Cameron Jones and Benjamin Bergen
- 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, Nope, the Turing Test has not been solved
ELIZA (built in 1965-1966) beat GPT 3.5.
- Where
- paragraph 3
- Published
- 2023-12-09
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-nope-the-turing-test-has-not-been-dd3f9e-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-nope-the-turing-test-has-not-been-dd3f9e-a1.