“Moreover, OpenAI (like every other generative AI model developer) is incapable of solving the critical flaw with ChatGPT, namely its tendency to hallucinate — where it asserts something to be true, when it isn’t. This makes it a non-starter for most business customers, where (obviously) what you write has to be true.”
Our reading:
OpenAI, like every other generative AI model developer, is incapable of solving ChatGPT's tendency to hallucinate, and this makes it a non-starter for most business customers.
Assessment: Unresolved
What happened
Does the evidence support the cause?
Outcome: Unresolved
We checked, and the evidence so far doesn’t settle it.
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
The June 19, 2024 Nature paper reports persistent false outputs and only partial success from truthfulness training [e1]. It does not establish that practical solutions are impossible [e1]. AICPA & CIMA reported on December 12, 2024 that 36% of surveyed executives said their companies were at least experimenting with generative AI [e2]. That finding does not measure rejection caused by hallucinations among most potential business customers [e2]. Neither the technical impossibility nor the claimed majority rejection is established by these items [e1, e2].
Caveats
- The technical evidence establishes neither a general solution nor a general impossibility.
- The retrieved survey headlines lack methods needed to assess representativeness.
- The evidence does not define the potential customer population or measure hallucination-caused rejection.
Evidence (2 items)
Context only: Nature,
only partially successful8.
Why it matters: The paper documents persistent factual errors and proposes a detection method. Incomplete mitigation does not establish that practical solutions are impossible or quantify business rejection.
- Where
- Abstract
- Publisher
- Nature
- Kind of source
- Peer-reviewed paper
- About events on
- Published
Context only: AICPA & CIMA,
at least experimenting
Why it matters: Business experimentation provides limited counterweight to the commercial dismissal. It neither establishes majority adoption nor identifies hallucinations as the cause of non-adoption.
- Publisher
- AICPA & CIMA
- Kind of source
- Independent measurement
- About events on
- Published
How we judged it
- Kind of claim
- Causal claim
- Deadline
- None given. We don’t add one.
Conditions
- most business customers (their words)
Where we had to interpret
- The plain reading is that factual unreliability cannot be practically solved and consequently prevents adoption by more than half of potential business customers.
- Solving means making useful general-purpose factual responses reliably truthful; neither a small benchmark improvement nor a system that refuses every question settles this.
- The commercial assertion concerns rejection caused by hallucinations, not merely concern about accuracy.
- This is assessed as a causal claim, not a forecast with an invented deadline or a historical adoption-rate estimate.
Right if: Evidence establishes the general technical limitation and representative business research shows that hallucinations cause a majority of potential business customers to reject adoption.
Wrong if: A reproducible general solution defeats the technical assertion, or representative evidence establishes that hallucinations do not make adoption a non-starter for most business customers.
We wrote this test after the outcome was known.
What would settle it
Reproducible research measuring false assertions, abstention and useful coverage, together with representative research on hallucination-caused rejection. Support requires both the general limitation and majority rejection. A general practical solution or representative evidence ruling out majority rejection would contradict the claim.
We found 49 claims in this post. See them all.
Full record
Source and dates
Where's Your Ed At, There Is No AI Revolution, paragraph 51.
- 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 2025-02-24.
Where it was said (1)
Said once in the posts we read.
: Where's Your Ed At, There Is No AI Revolution
Moreover, OpenAI (like every other generative AI model developer) is incapable of solving the critical flaw with ChatGPT, namely its tendency to hallucinate — where it asserts something to be true, when it isn’t. This makes it a non-starter for most business customers, where (obviously) what you write has to be true.
- Where
- paragraph 51
- Published
- 2025-02-24
- We read it
- , live page
- Access
- full text, public
History
Versions
- Current: Unresolved. Evidence up to 2026-10-28.
What happened to this record
PublishedBy us
Published after both AI reviews and the final check passed.
Held backBy us
The final check failed: the write-up said more than the evidence showed.