“Any one example can be filtered out, but there are too many examples, and current systems are too erratic to reliably get things right.”
Our reading, the sentence that was checked:
Because there are too many examples and current systems are too erratic, guardrails cannot reliably get things right even though any single example can be filtered.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, DALL-E’s New Guardrails: Fast, Furious, and Far from Airtight, paragraph 12.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-dall-es-new-guardrails-fast-furious-773ec1-a1
We recorded 9 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 2024-01-10. 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
- Any one example can be filtered out
- too many examples
- current systems are too erratic
- reliably get things right
Our interpretation
- The claim attributes unreliable safeguards to the diversity of possible inputs and inconsistent system responses, despite the ability to block a known example.
- Reliability means consistently enforcing intended restrictions across varied prompts and repeated trials. No numerical reliability threshold is supplied.
- Current systems refers to systems around the post date. I use December 2023 through January 10, 2024 as the recent observation window.
Counts as supported if: Controlled evaluations show that prompt diversity and response variability cause continuing safeguard failures despite successful filtering of individual examples.
Counts as contradicted if: Controlled evaluations establish reliable generalization under diverse prompts and repeated trials, or show that the asserted causal factors do not explain the failures.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Does the evidence support the cause?
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
The sources establish vulnerabilities and show that broader defenses can reduce them without eliminating them. They do not settle the claimed explanation that excessive example diversity and erratic responses cause the unreliability. Treating jailbreak success alone as proof would collapse a causal claim into a weaker observation.
Limitations
- The controlled defense study is contextual because its experiments predate the target period.
- The causal terms and reliability threshold remain underspecified.
Evidence (2 items)
Context only: National Institute of Standards and Technology,
bad or toxic information
Why it matters here: NIST's contemporary assessment recognizes vulnerabilities to carefully constructed prompts. It does not experimentally separate prompt diversity and response inconsistency as causes of unreliable safeguards.
- Publisher
- National Institute of Standards and Technology
- Kind of source
- Secondary summary
- Describes events of
- Public from
Context only: Nature Machine Intelligence; author-hosted copy,
from 67.21% to 19.34%.
Why it matters here: A general defense substantially reduced attacks while leaving residual failures. This qualifies any reading that safeguards could only block individual examples, but the experiments predate the target window and do not isolate Marcus's full causal explanation.
- Where
- Page 1, abstract
- Publisher
- Nature Machine Intelligence; author-hosted copy
- Kind of source
- Peer-reviewed paper
- Describes events of
- Public from
What would settle it
Contemporaneous controlled tests comparing individual fixes with general defenses, varying prompt diversity and repeated sampling, with a stated reliability criterion.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, DALL-E’s New Guardrails: Fast, Furious, and Far from Airtight
Any one example can be filtered out, but there are too many examples, and current systems are too erratic to reliably get things right.
- Where
- paragraph 12
- Published
- 2024-01-10
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-dall-es-new-guardrails-fast-furious-773ec1-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-dall-es-new-guardrails-fast-furious-773ec1-a1.