“But here’s the thing: although a lot of the literature equates artificial intelligence risk with the risk of superintelligence or artificial general intelligence, you don’t have to be superintelligent to create serious problems.”
Our reading, the sentence that was checked:
AI does not need to be superintelligent to create serious problems.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, AI risk ≠ AGI risk, paragraph 28.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-ai-risk-agi-risk-c1e258-a2
We recorded 8 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 2023-03-28. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Causal claim
- 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
- you don’t have to be superintelligent
- to create serious problems
Our interpretation
- This is a claim that serious harm is possible without superintelligence, not a prediction of catastrophe.
- The ordinary reading of serious problems includes death, serious injury, or substantial material harm.
- A documented contribution by a limited-purpose AI system is sufficient. AI need not be the sole cause.
- The introductory characterization of the literature is outside the supplied normalized proposition.
- A specialized driving system with documented perception and response limitations is treated as below superintelligence.
Counts as supported if: A credible investigation establishes that a below-superintelligent AI system contributed materially to serious harm.
Counts as contradicted if: Evidence establishes that the proposed examples did not involve AI or did not causally contribute to serious harm, together with a defensible basis for finding that such harm requires superintelligence. Failure to find an example alone would not establish contradiction.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Does the evidence support the cause?
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 NTSB investigation of the March 18, 2018 fatal collision documents failures in Uber’s automated driving system: it misclassified the pedestrian, failed to predict her path, and precluded emergency braking [e1]. These limitations support a material contribution to serious harm by a specialized system treated as below superintelligence under the contract [e1]. The NTSB identified the distracted operator as the probable cause and also identified organizational and oversight failures as contributing factors [e1].
Limitations
- This incident establishes that serious harm is possible, not how frequently it occurs [e1].
- The investigation does not establish AI as the sole cause or demonstrate civilization-scale harm [e1].
- Treating this system as below superintelligence is the contract’s interpretation of its documented capabilities and limitations, not an NTSB classification.
Evidence (1 item)
Supports the claim: National Transportation Safety Board,
The pedestrian died in the crash.
Why it matters here: The investigation describes a fatal collision while Uber's developmental driving system controlled the vehicle; it failed to classify the pedestrian correctly or predict her path, and its design suppressed emergency braking. This supports a material contribution from limited AI, while the report identifies the distracted operator as the probable cause and also identifies organizational and oversight failures.
- Publisher
- National Transportation Safety Board
- Kind of source
- Independent measurement
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, AI risk ≠ AGI risk
But here’s the thing: although a lot of the literature equates artificial intelligence risk with the risk of superintelligence or artificial general intelligence, you don’t have to be superintelligent to create serious problems.
- Where
- paragraph 28
- Published
- 2023-03-28
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-ai-risk-agi-risk-c1e258-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.