The AI Act contains some specific provisions dealing with the possible use of artificial intelligence for discriminatory purposes or in discriminatory ways, in the context of the European Union. The AI Act also regulates generative AI models. However, these two respective sets of rules have little in common: provisions concerning non-discrimination tend not to cover generative AI, and generative AI rules tend not to cover discrimination. Based on this analysis, the Chapter considers what is currently the Eu legal framework on discriminatory output of generative AI models, and concludes that those expressions that are already prohibited by anti-discrimination law certainly remain prohibited after the approval of the AI Act, while discriminatory content that is not covered by Eu non-discrimination legislation will remain lawful. For the moment, the AI Act has not brought any particularly relevant innovation on this specific matter, but the picture might change in the future.

Threats are not protected speech, but defining what constitutes a threat has been problematic, particularly when it comes to online speech. We start with a look at threats against the US president, beginning with a 1798 prosecution for threatening John Adams, and leading up to the passage in 1917 of the first federal legislation against threatening the president. We will look at WW I-era prosecutions for threatening the president, leading up to a 1969 Supreme Court decision, Watts v. US, distinguishing true threats from protected political speech. And we conclude with two cases of online threats: the prosecution of Anthony Elonis for posting threats on Facebook, and a case where two tourists were denied entry to the US because of joking tweets that were treated as threats by US Border Agents. We conclude that threats, like obscenity, remain unprotected speech, but defining what is and what is not a threat in any particular case remains a problematic, subjective decision.