“Businesses are finally finding this out, too.”
Our reading:
Businesses are discovering that generative AI tools have major limitations and are riddled with errors.
Assessment: Supported
What happened
Was it true for the period described?
Outcome: Supported
The evidence matches what they said.
Judged on a plain reading. They named no measure, so we used the reading a typical informed reader would take. It is spelled out under “How we judged it”.
Evidence up to . Checked .
Why
McKinsey's contemporary survey documents businesses experiencing consequential inaccuracies and growing recognition of that risk. This supports the ordinary reading that businesses were discovering practical limitations. The same report records benefits, so it does not support the surrounding suggestion that generative AI was generally useless.
Caveats
- The evidence does not establish that most businesses experienced harm or that most outputs were erroneous.
- The primary report was read through an externally hosted copy; available_at refers to the original report's publication.
- The business-benefit findings in e1 limit the broader negative interpretation but do not contradict recognition of errors.
Evidence (1 item)
Supports the claim: McKinsey & Company; report copy hosted by ABBYY,
negative consequences
Why it matters: At the close of February 22–March 5 fieldwork, 23% of respondents at organizations adopting generative AI reported adverse consequences from inaccuracies. The report also records increased recognition of inaccuracy risk compared with its previous survey, while separately reporting business benefits.
- Publisher
- McKinsey & Company; report copy hosted by ABBYY
- Kind of source
- Independent measurement
- About events on
- Published
How we judged it
- Kind of claim
- Observation
- Deadline
- None given. We don’t add one.
Where we had to interpret
- The ordinary reading is that businesses using generative AI were encountering substantial errors and recognizing its limitations. It does not require every business to agree or establish that the technology has no value.
- For this ongoing observation, use early-to-mid-2024 business experience reported before the June 28 post: January 1 through June 28, 2024.
- Read 'finally' as increasing practical recognition, rather than a claim that no business previously understood these risks.
Right if: Contemporary business surveys or documented deployments show that businesses encountered consequential inaccuracies and increasingly recognized limitations.
Wrong if: Representative contemporary evidence shows that substantial inaccuracies were uncommon and that businesses were not encountering the described limitations.
We wrote this test after the outcome was known.
We found 12 claims in this post. See them all.
Full record
Source and dates
Marcus on AI, The need for a President that speaks AI natively, paragraph 5.
- Said
- We read the source
- live page, full text, public
- Evidence up to
- Checked
We read this wording on 2026-10-28. The post may have been edited after 2024-06-28.
Where it was said (1)
Said once in the posts we read.
: Marcus on AI, The need for a President that speaks AI natively
Businesses are finally finding this out, too.
- Where
- paragraph 5
- Published
- 2024-06-28
- We read it
- , live page
- Access
- full text, public
History
Versions
- Current: Supported. Evidence up to 2026-10-28.
What happened to this record
PublishedBy us
Published after both AI reviews and the final check passed.