OpenAI’s ChatGPT automatically blocks conversations containing certain individuals’ names, according to reporting by Benj Edwards for Ars Technica. The blocking mechanism appears to be a response to defamation concerns and legal threats, with confirmed blocks on names including Brian Hood, Jonathan Turley, and Jonathan Zittrain. The article explains that Hood’s case, involving false claims about imprisonment for bribery, led to one of the first such blocks after a settled lawsuit. These hard-coded filters can cause practical problems, including potential exploitation for adversarial attacks and difficulties for innocent users who share blocked names. The blocks affect only ChatGPT’s main interface, not OpenAI’s API or Playground, and represent an early attempt to prevent the AI from generating false, potentially defamatory statements about specific individuals.