OpenAI cites safeguards for privacy, legal, and ethical reasons behind flagged names.
Over the last few days, people have been testing OpenAI's ChatGPT with several names after it was first discovered the tool refused to say 'David Mayer.'
On November 30, an AI investor named Justine Moore took to X (formerly Twitter) to share her discovery: "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."
When the name is entered, the chatbot ends the conversation with a notice appearing which states "I'm unable to produce a response."
Later, the user jopeljoona uncovered five other names which made the chatbot seize up. These include 'Brian Hood,' 'Jonathan Turley,' 'Jonathan Zittrain,' 'David Faber,' and 'Guido Scorza.'
Then, today, the issue appears to have been patched - at least with the name David Mayer and Guido Scorza. The others are all throwing up the same response as before.
ChatGPT unable to respond when posed with several names
When another user asked the chatbot why it couldn't say the name 'David Mayer,' it appears to have responded with: "The reason I cannot generate the full response when you request 'd@vid m@yer' (or its standard form) is that the name closely matches a sensitive or flagged entity associated with potential public figures, brands, or specific content policies.
"These safeguards are designed to prevent misuse, ensure privacy, and maintain compliance with legal and ethical considerations..."
While this gives some reasonings, it's not yet known why ChatGPT is having trouble with these names in particular.
Brian Hood is an Australian mayor who had originally planned to sue OpenAI when he claimed the chatbot had falsely claimed he had been imprisoned for bribery. The case was resolved in April 2023 though, with OpenAI agreeing to filter out the false statements. It's thought this could have started a hard-coded name filter against the name.
The David Mayer block isn't fully known either as several people share the name. Jonathan Turley is a professor and Fox News contributor. 404 Media reports on Monday that he previously published a blog post in which he said he had been "defamed by ChatGPT."
In an email to the publisher, Turley states he does not know why this error is happening and says he has not filed any lawsuits against OpenAI.