“But time has shown that its results have held strong.”
Our reading, the sentence that was checked:
The MIT finding that only 5% of companies get a return on generative AI investment has held up.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Three years on, ChatGPT still isn't what it was cracked up to be – and it probably never will be, paragraph 13.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-three-years-on-chatgpt-still-isnt-d75209-a1
We recorded 30 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2025-11-29. 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)
From the speaker
- its results
- held strong
- only 5% of companies were getting a return on generative AI investment
- McKinsey
- the results weren’t much different
Our interpretation
- Its results refers to the MIT NANDA finding described immediately before this sentence.
- Held strong means subsequent evidence corroborated the approximately 5% finding, rather than merely showing that adoption remained difficult.
- The target is the interval from the report's stated July 2025 date through the post. Survey fieldwork dates must be distinguished from publication dates.
- A share receiving any return, a share achieving measurable profit-and-loss impact, and a share classified as high performers are not interchangeable.
- The metric is a prevalence estimate across organizations, not an identified company's financial accounts.
Counts as supported if: Subsequent evidence with sufficiently comparable organizations, denominators, and return definitions corroborates an approximately 5% share obtaining the stated benefit.
Counts as contradicted if: Subsequent comparable evidence establishes a substantially different share, or directly demonstrates that the original finding did not hold.
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
The retrieved evidence does not establish that the MIT finding held up on a comparable measure. Wharton's later publication reports positive returns for roughly three-quarters of leaders, which challenges Marcus's broad wording. But its population and return definition differ from MIT's pilot and P&L measures. That prevents treating the difference as a clean numerical refutation. The cited McKinsey comparison could not be verified through readable source text.
Limitations
- Similar difficulties scaling AI do not independently corroborate a 5% return prevalence.
- Wharton's survey is material contrary context, not a matched replication.
- Access failures limited verification of the comparison Marcus expressly cited.
Evidence (2 items)
Context only: MIT NANDA report authors; copy hosted by MLQ,
integrated AI pilots
Why it matters here: This is the original baseline, based on January–June research, not subsequent corroboration. Its formulation concerns integrated pilots and measurable P&L impact, which complicates comparison with a claim about all companies receiving any return.
- Publisher
- MIT NANDA report authors; copy hosted by MLQ
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: Wharton Human-AI Research and GBK Collective,
Three out of four leaders see positive returns on Gen AI investments.
Why it matters here: The later publication challenges the broad characterization that almost nobody receives returns. Its June 26–July 11 survey covered large US enterprises and used a broader, self-reported ROI concept, so it cannot directly settle the MIT P&L finding.
- Publisher
- Wharton Human-AI Research and GBK Collective
- Kind of source
- Independent measurement
- Describes events of
- Public from
What would settle it
A readable subsequent study using a comparable population, denominator, and financial-return definition, or sufficiently detailed survey data to reconcile those differences.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Three years on, ChatGPT still isn't what it was cracked up to be – and it probably never will be
But time has shown that its results have held strong.
- Where
- paragraph 13
- Published
- 2025-11-29
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-three-years-on-chatgpt-still-isnt-d75209-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-three-years-on-chatgpt-still-isnt-d75209-a1.