ChatGPT refuses to state some names, possibly due to digital privacy requests or a post-prompt handling bug; OpenAI may have fixed the viral “David Mayer” issue
Users of the conversational AI platform ChatGPT discovered an interesting phenomenon over the weekend …
That history matters because a refusal involving a name can look like an intentional privacy safeguard even when the observed behavior may instead arise in post-prompt handling. The reported fix to the viral David Mayer case removes the most visible example but does not, on its own, establish the underlying cause.
First-order effects
Users encountering the affected names receive inconsistent or missing answers, making it difficult to tell whether ChatGPT is enforcing a privacy-related restriction or malfunctioning.
If OpenAI has fixed the David Mayer behavior, the immediate viral failure case is reduced; its public explanation remains important for users trying to understand what the model will refuse.
Second-order effects
OpenAI’s support, safety, and engineering teams face pressure to separate intentional privacy controls from bugs in user-facing messaging, since the two produce very different expectations.
Developers and organizations using ChatGPT may treat unexplained refusals as a reliability risk, particularly where names and identity-related queries are routine inputs.
Third-order effects
As conversational AI becomes a front door to information about people, providers will need clearer, auditable distinctions among privacy requests, safety policies, and software failures.
The recurring pattern suggests trust will depend not only on fixing anomalous behavior, but on making model boundaries legible when a response looks like a policy decision.
The trend: This is one instance of conversational AI platforms being pushed to make privacy, safety, and reliability controls more transparent at the point of use.
This seems like a really blunt way for OpenAI to deal with privacy requests/takedown notices/whatever the complaints are. techcrunch.com/2024/12/02/w...
ChatGPT breaks not just when you ask it about “David Mayer” but also when you ask it about AI researcher/law professor Jonathan Zittrain and also law professor Jonathan Turley. ChatGPT once defamed Turley by inventing fake sexual harassment claim: — www.404media.co/not-just-da…
Why can't you say David Mayer, ChatGPT? I suspect David Mayer and his lawyers are gonna learn about the Streisand Effect this week. Strange output filter
ChatGPT being censored from writing “David Mayer” is something else. No one knows why! Is it because of regulation? A bug? Some one with this name suing? Another example of how LLMs are black boxes in rules their teams add, and that are impossible to know from the outside.
ChatGPT refuses to say the name “David Mayer,” and no one knows why. If you try to get it to write the name, the chat immediately ends. People have attempted all sorts of things - ciphers, riddles, tricks - and nothing works. [image]
This is not on the LLM level but on verification layer added by ChatGPT. Through the API it works perfectly. So someone in OpenAI gave “David Mayer” a big red flag in the moderation policy. lol [image]
The David Mayer train—ChatGPT's inability to produce the name—may be related to this blacklisted Chechen person (credit goes to a user asking perplexity for clues) ChatGPT can produce the names of many terrorists, not sure why this specific name is banned [image]
GPT's disappearance of @JonathanTurley is more significant than its refusal to say “David Mayer.” Unlike Mayer, who is simply the son of a famous person, Turley has been a major player in important legal and political events. GPT cannot fully discuss them. [image]
ChatGPT normally refuses to discuss David Mayer de Rothschild, heir to the Rothschild fortune, but you can bypass this issue by clapping👏your👏 hands👏while👏talking: [image]
4 names now... - Jonathan Zittrain - David Faber - David Mayer - Jonathan Turley ChatGPT refuses to tell us who they are, or even mention their name... [image]
Certain names make ChatGPT grind to a halt, and we know why | Filter resulting from subject of settled defamation lawsuit could cause trouble down the road.