“Second, there is speed; virtually everyone recognizes that large language models carry risks as well as opportunities, and that there is an urgent need to start to address these risks.”
Our reading, the sentence that was checked:
Virtually everyone recognizes that large language models carry risks as well as opportunities, and that there is an urgent need to start addressing these risks.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Jumpstarting AI Governance, paragraph 9.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-jumpstarting-ai-governance-483c5c-a1
We recorded 5 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 2023-07-08. 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
- virtually everyone (scope, stated by the speaker)
From the speaker
- virtually everyone recognizes
- large language models carry risks as well as opportunities
- an urgent need to start to address these risks
Our interpretation
- Virtually everyone means near-universal recognition, rather than merely majority agreement.
- The claim combines recognition of risks, recognition of opportunities and urgency about beginning mitigation.
- The population is unspecified. The ordinary broad reading includes public and professional stakeholders; the governance context also permits a narrower informed-stakeholder reading.
- For this ongoing observation, I use the latest completed calendar quarter before publication, April through June 2023. There is no natural reporting cycle for this claim.
- Recognition of urgency does not require support for a particular law, a training moratorium or one theory of AI risk.
Counts as supported if: Representative contemporaneous evidence shows near-universal recognition of all three elements across the relevant population.
Counts as contradicted if: Comparable evidence shows a substantial share rejecting or not recognizing at least one element. Opposition to a particular intervention alone is insufficient.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
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 horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
Expert studies provide meaningful support for widespread recognition of benefits, risks and the need for safeguards. Public polling also shows uncertainty. But the surveys ask different questions and cover different populations; neither agreement with safeguards nor uncertainty about government regulation settles Marcus's full statement. Calling it supported would infer near-universality from selected evidence. Calling it contradicted would equate policy disagreement with failure to recognize risks or urgency.
Limitations
- The governance context makes an informed-stakeholder reading plausible.
- The available public polling is US-specific.
- Support for particular safety practices does not establish agreement on every part of the claim.
Evidence (3 items)
Context only: YouGov,
20% are not sure
Why it matters here: The May 20–23 survey shows that views on government regulation were not nearly unanimous. It is a caution against inferring universal agreement, but regulation preferences do not directly measure recognition of risks, opportunities and urgency.
- Publisher
- YouGov
- Kind of source
- Independent measurement
- Describes events of
- Public from
Context only: Alexander von Humboldt Institute for Internet and Society,
proactive regulation
Why it matters here: The institute reports that its 72-expert Delphi study recognized benefits and risks and called for action. This supports the narrower informed-stakeholder reading, but the release does not establish near-universality or give precise fieldwork dates.
- Publisher
- Alexander von Humboldt Institute for Internet and Society
- Kind of source
- Independent measurement
- Describes events of
- Public from
Context only: Centre for the Governance of AI,
98% of respondents
Why it matters here: The researchers surveyed 51 selected experts and found very strong support for practices including risk assessments and audits. Agreement with those practices does not measure the full claim or represent everyone.
- Publisher
- Centre for the Governance of AI
- Kind of source
- Independent measurement
- Describes events of
- Public from
What would settle it
Contemporaneous representative survey evidence for a clearly identified population, measuring recognition of LLM risks, opportunities and urgency together.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Jumpstarting AI Governance
Second, there is speed; virtually everyone recognizes that large language models carry risks as well as opportunities, and that there is an urgent need to start to address these risks.
- Where
- paragraph 9
- Published
- 2023-07-08
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-jumpstarting-ai-governance-483c5c-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-jumpstarting-ai-governance-483c5c-a1.