The company Meta (Facebook, Instagram, WhatsApp) has announced that it will advocate for a ban on the use of artificial intelligence tools for political advertising and communication with voters in all countries where their users are located. Those tools, by the way, help advertisers reach customers by creating quality ads and the analytics that go with them, with the fact that Meta is now trying to limit it to those who sell drugs, financial services, real estate, and pay special attention to elections.
Artificial intelligence could have both a positive and a negative role in political campaigns, with the fear of the negative outweighing what would be useful to both politicians and voters.
In a positive sense, artificial intelligence could help when drafting laws, but also when analyzing proposed laws. It would save time and indicate what is very important to a political option. Assisting artificial intelligence in collecting donations and creating political messages could also be considered positive to some extent. Let's say, when asked to propose a slogan for an unspecified party in the upcoming elections in Serbia, chatGPT came up with the following: "Together for a better Serbia." Security. Prosperity. Unity.” The slogan was followed by a short explanation of why these terms are currently key in Serbia, that is, attractive to voters.
A chat bot could easily define and explain the party's policy, especially for different social groups and ages. Then some other AI tool could market them to exactly those it was created for. Virtually without any creative human influence except for the content the bot was trained on.
It is not punishable, not even too immoral, the question is whether it can be limited. A much bigger problem is the negative messages and fake news that a chatbot can create and adapt to its target audience. Although bots are designed to not write negatively, these limitations can be circumvented relatively easily.
That's why experts in Internet politics recommend users to be very careful during the election campaign when it comes to social networks. The most radical advice is to not use them at all until the election, none of them, because there is every chance that there are messages created to our liking lurking there.
Another tip is to be paranoid, i.e. don't believe anything you see or read online. The first impression can significantly affect your attitude and therefore everything, they say, should be treated as untruth or deception. And then commit to verifying the information from multiple sources until you determine what is really true.
When all this is added up, we arrive at a somewhat frightening scenario. A capable user of artificial intelligence could, with its help, devise an ideology, form a party, enroll people whom he would catch on social networks with the help of algorithms and a combination of fake news and manipulation, and then attract voters in a similar way. Until a year ago this would have been a good basis for a science fiction novel. Today, it is not even science fiction that instead of that human, the user, all this will be done by artificial intelligence, while people would only serve as voters.