ChatGPT Crashes If You Mention the Name “David Mayer”

We still have no idea why.

David Who?

Netizens were baffled to find that OpenAI’s ChatGPT refused to acknowledge the existence of the name “David Mayer.”

Ask the chatbot about the name and it gives a puzzling reply: “I’m unable to produce a response.”

Then the chat abruptly ends, forcing users to open a new window to keep using the assistant.

Frustratingly, OpenAI has yet to comment on the matter and questions are still hard to come by. The most prominent personality who goes by that name is David Mayer de Rothschild, a British adventurer and environmentalist who presumably doesn’t have any bones to pick with an AI chatbot.

He also happens to be the heir of the Rothschild banking dynasty, which has spawned plenty of far-fetched conspiracy theories.

The incident highlights just how little we know about how these large language models operate, what kind of content they’ve been trained on, and what limits are imposed by their makers.

We’ve seen plenty of examples of AI chatbots hallucinating facts, failing at simple tasks, and even encouraging dangerous behavior — meaning that the tool getting hung up on a simple name isn’t even the most unusual behavior we’ve come across.

Shmavid Shmayer

As we wait for official word from OpenAI — Futurism has reached out for comment — many social media users have been busy digging around for some answers.

For instance, AI investor Justine Moore found at least six more names that triggered a similar response, including several individuals who have filed General Data Protection Regulation (GDPR) “right to be forgotten” requests in the EU. In other words, the quirk may have to do with their efforts to have their online presence erased.

Others have suggested that the name David Mayer may have been blocked due to the existence of a Chechen ISIS member who used the name as an alias.

As 404 Media reports, ChatGPT also breaks when it’s asked about two law professors named Jonathan Zittrain and Jonathan Turley.

Intriguingly, Zittrain wrote a piece in The Atlantic about the need to “control AI agents now.”

One user even changed their name to David Mayer in ChatGPT, which led to the tool refusing to acknowledge their new name.

Is the name being blocked because of OpenAI’s guardrails, which were designed to protect individuals’ privacy? Did David Mayer — whoever he may be — file legal documents with the AI company?

It’s a mystifying bug that underlines how much of a black box ChatGPT really is. And it underscores a constant lesson with AI products: don’t take any of their answers at face value.

More on ChatGPT: Lawyer Submits Anti-AI Document That Appears to Have Been Created Using AI