As Mr. Kingsley said, as soon as the message is controlled by artificial intelligence and the author is identified, if it's satire, it can still be acceptable. We can't prevent or even limit comedians from enjoying freedom of expression.
On the other hand, if it really isn't humour and people are trying to mislead the public and systematically give false information, voters can file complaints, and there will be an investigation. If we're able to establish beyond a reasonable doubt that the person acted voluntarily, there will be sanctions. It isn't a perfect solution, but we're in the early days.
I'll compare it to computer viruses. Viruses arrived and took up space, but there have always been antiviruses to counter them, so we also have to trust in organizations. Artificial intelligence is going to be harmful, but it will also be able to help find the causes of disinformation.