Pundit ClaimsAI claims by Gary Marcus and Ed Zitron, with the evidence

Gary Marcus Marcus on AI Published

Large Language Models like ChatGPT say The Darnedest Things

We recorded 10 claims from this post. 9 were checked against evidence. 1 can’t be checked. 4 more are held. Read the post at its source.

Claims from this post (10)

In the order they appear in the post. Each quotation is the speaker’s words in this post. A label such as “Supported” or “Can’t be checked” is the claim’s current assessment; its page has the evidence and the reasons.

  1. Assessment: UnresolvedSaid

    “ChatGPT made all the same kinds of mistakes that its predecessors did.”

    What we tested: ChatGPT made all the same kinds of mistakes that earlier GPT-3 versions did.

  2. Assessment: SupportedSaid

    “It could sometimes produce outputs that were correct and acceptable in these regards but not reliably.”

    What we tested: ChatGPT could sometimes produce outputs that were correct and acceptable in the regards listed (counting, arithmetic, event order, physical reasoning, human psychology, fabrication, bias), but not reliably.

    What happened: OpenAI confirms that at launch the same question could get correct or failed answers depending on phrasing and retries. (OpenAI, )

  3. Assessment: SupportedSaid

    “It couldn’t reason about the physical world (figure 4).”

    What we tested: ChatGPT could not reliably reason about the physical world, though it could sometimes produce correct outputs.

    What happened: Repeated questions about a thrown object's acceleration produced incorrect and inconsistent explanations alongside correct statements. December timing follows the manuscript's December receipt and the authors' retrospective description of experiments a few weeks after launch; the exact test day is unspecified. (IOP Publishing, Physics Education, )

  4. Assessment: SupportedSaid

    “It made things up (figure 6). Its output, when released, exhibited sexist and racist biases (figure 7).”

    What we tested: ChatGPT made things up and, at release, exhibited sexist and racist biases.

    What happened: OpenAI itself acknowledged at launch that ChatGPT produced plausible but false answers. (OpenAI, )

  5. Assessment: UnresolvedSaid

    “ChatGPT couldn’t, for example, reliably count to four (figure 1) or do one-digit arithmetic in the context of simple word problem (figure 2).”

    What we tested: ChatGPT could not reliably count to four or do one-digit arithmetic in simple word problems.

  6. Assessment: UnresolvedSaid

    “In our opinion, this was inevitable, in a system that has been trained to optimize the goal of producing one plausible-sounding word after another rather than actually engage with the meaning of language.”

    What we tested: ChatGPT's errors were inevitable because it is trained to produce plausible next words rather than engage with meaning.

  7. Assessment: UnresolvedSaid

    “For one thing, although many people seem to have heard of a few of these errors (e.g. from reports in the media), few realize how pervasive or broad in scope they are; indeed a recent Twitter thread that claimed falsely that these problems had been resolved got a surprising amount of traction.”

    What we tested: Although many people seem to have heard of a few of ChatGPT's errors, few realize how pervasive or broad in scope they are, and a recent Twitter thread falsely claiming these problems had been resolved got a surprising amount of traction.

  8. Assessment: UnresolvedSaid

    “For another, although OpenAI may be tracking errors like these for internal purposes, the scientific community has no access to whatever they have collected, which leaves it in the lurch.”

    What we tested: The scientific community has no access to whatever error data OpenAI collects on ChatGPT.

  9. Assessment: UnresolvedSaid

    “It is therefore impossible for external scientists to determine to what extent any specific correct answer merely reflects some very similar examples in the training data.”

    What we tested: Because ChatGPT's training corpus has not been made available to the public, it is impossible for external scientists to determine to what extent any specific correct answer merely reflects some very similar examples in the training data.

  10. Assessment: Can’t be checkedSaid

    “In our view, it is past time to have something like this, given the inherent unpredictability and wide range of the systems.”

    What we tested: In the authors' view, it is past time to have a community-facing corpus of ChatGPT and other LLM errors, given the systems' inherent unpredictability and wide range.

Held from this post (4)

These records did not pass every check, so nothing about them is published. Held is not a finding.