Slovakia’s Election Deepfakes Show AI Is a Danger to Democracy
wired.com
wired.com
The fake here was not subtle. It was not trying to nudge people away from a position. It said baldly "This candidate is a criminal and nobody should even consider voting for them."
In a system with any decency, the opponent should say "That is categorically false and I condemn whoever made this fake. The election is about the legitimate differences between us, but my opponent is a good person. They'd be a good in the office. I would be better, and you should vote for me, but not because you've been fed lies."
Instead we've spent decades reducing every election to an existential crisis. At least in the US -- I can't speak to other countries. Differences in policy or managerial style are ignored in favor of "they are trying to destroy the country".
Which is aggravatingly self-fulfilling. Once you've decided that your opponents are enemies of the state, it justifies every response, including unethical or even illegal ones.
Deepfakes exacerbate that, but they seem like the inevitable consequence of a population who was carefully prepared to believe outright, transparent lies. You're right that AI will addle democracy further, but only because we were already addled.
There are much subtler ways to use deepfakes to persuade rather than harass. But it sounds like we'll never get to those, since people are perfectly willing to accept the crude ones.