In January, the Alberta Party, a centrist provincial party, posted a video on its official Instagram account. The since deleted video, which is still available on their Twitter feed, was of a man in a blue sweater facing the camera, with the Calgary skyline behind him. But something about him was off. “Does Alberta need a third political party? It depends on whether or not you’re happy with your current choices,” the man stated flatly, before suggesting Albertans seek “another option” on election day. His mouth didn’t form vowels properly. His face was strangely unmoving. And then there was his voice—it didn’t sound fully human.
That’s because it wasn’t. The video was AI generated. In a since deleted post, the Alberta Party clarified that “the face is an AI, but the words were written by a real person.” It was an unsettling episode, not unlike one from the hit techno-dystopian series Black Mirror in which a blue animated bear, Waldo, voiced by a comedian, runs as an anti-establishment candidate in a UK by-election.
“ . . . wherever the people are well informed they can be trusted with their own government,” Thomas Jefferson wrote in the winter of 1789. In other words, in a democracy, being informed grants people agency, and having agency gives us a voice. AI has the potential to challenge both of these critical assumptions embedded for hundreds of years at the heart of our understanding of democracy. What happens when the loudest, or most influential, voice in politics is one created by a computer? And what happens when that computer-made voice also looks and sounds like a human? If we’re lucky, AI-generated or AI-manipulated audio and video will cause only brief moments of isolated confusion in our political sphere. If we’re not, there’s a risk it could upend politics—and society—permanently.
Read more: AI and Politics: How Will We Know What—and Who—Is Real?