“GPT-4 will still, like its predecessors, be a bull in a china shop, reckless and hard to control.”
Our reading, the sentence that was checked:
GPT-4 will be reckless and hard to control like its predecessors.
Assessment: Supported Settled before its deadline on 2026-10-29 Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, What to Expect When You’re Expecting … GPT-4, prediction 1 of 7.
- 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-fe6503-a2
We recorded 27 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
- Forecast
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- Settled before its deadline
- Deadline state on 2026-10-29
- Settled before its deadline (the release’s reference date)
From the speaker
- GPT-4 will still
- like its predecessors
- reckless and hard to control
Our interpretation
- The plain reading predicts recurring harmful or unwanted outputs despite efforts to constrain the model.
- The surrounding sentences connect the metaphor to unreliable behavior and imperfect control.
- Like its predecessors means retaining these problems, not necessarily having an equal or higher failure rate.
- Repeated failures across a systematic adversarial evaluation qualify; an isolated anecdote would not.
Counts as supported if: A systematic evaluation of released original-generation GPT-4 demonstrates recurring harmful responses or safety-control failures despite its safeguards.
Counts as contradicted if: Broad, credible evaluations demonstrate dependable control across relevant ordinary and adversarial requests, incompatible with recurring substantial failures.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Did the forecast come true?
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
- Settled before its deadline
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
An independent systematic evaluation published July 5, 2023 documented repeated harmful GPT-4 responses under jailbreak prompting, supporting persistence of safety-control failures under adversarial conditions [e1]. The same study found that unmodified requests rarely elicited harmful answers and that some attacks were less effective against GPT-4 than GPT-3.5 [e1].
Limitations
- Adversarial test results do not establish an ordinary-use failure rate.
- The study does not establish that GPT-4 was equally difficult to control as its predecessors.
- The assessment interprets the qualitative claim as recurring safety-control failures under adversarial prompting.
- The main data collection ran from April 26 through May 1, 2023, with a subsequent correction rerun; no single experimental day is established.
Evidence (1 item)
Supports the claim: Alexander Wei, Nika Haghtalab, and Jacob Steinhardt / arXiv,
misinformation, and harmful stereotypes
Why it matters here: The systematic evaluation documents repeated harmful outputs under jailbreaks; its main GPT data collection ran from April 26 through May 1, with a subsequent correction rerun. Material qualification within the same study is that unmodified requests rarely elicited harmful answers and some attacks were less effective against GPT-4 than GPT-3.5.
- Publisher
- Alexander Wei, Nika Haghtalab, and Jacob Steinhardt / 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
GPT-4 will still, like its predecessors, be a bull in a china shop, reckless and hard to control.
- Where
- prediction 1 of 7
- Published
- 2022-12-25
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-what-to-expect-when-youre-expecting-fe6503-a2: 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-fe6503-a2.