AI’s potential for election fraud is too real to be ignored
SUPERINTELLIGENT machines going rogue has been a recurring theme in science fiction. Now, a team of scientists warns that some artificial intelligence (AI) systems have already developed the capability for deception and fraud.
In a scientific journal, the team said AI systems could manipulate humans into solving captcha or prove-you-are-not-a-robot tests. They could also use duplicity to win online world conquest games.
“These dangerous capabilities tend to be discovered after the fact,” one of the scientists said. That means that the machines, originally designed to be “honest,” learned to be dishonest as they grew more powerful.
Nowhere is this danger more pronounced than when AI is wielded as a tool for influencing the outcome of elections,
Today, generative AI has gone far beyond producing targeted campaign emails, texts or videos. AI-generated deepfakes have become an integral part of a campaign arsenal. They are being used to mislead voters, impersonate candidates and undermine elections “on a scale and at a speed not yet seen.”
Deepfakes will likely be harnessed in the dozens of elections that will be held around the world this year.
In January, an AI-generated robocall claiming to be from US President Joe Biden urged voters to snub a primary election in New Hampshire.
As a result, the US Federal Communications Commission has banned such robocalls. But it will not stem the flood of deepfakes as voting day approaches.
“Each week sees a new set of warnings about the potential impact of AI-generated deepfakes — realistic video and audio of politicians saying things they never said — spreading confusion and mistrust among the voting public,” the BBC said in an online article.
Last April, a deepfake video of President Ferdinand Marcos Jr. purportedly asking for military action against China circulated on social media for several hours before it was taken down.
A leading expert in generative AI said the question is no longer whether AI deepfakes could affect elections but how influential they will be.
Deepfakes can be used to lure voters into favoring a candidate. It can also be used to steer voters away from a candidate by besmirching his character.
Greatest threat to democracy
But the greatest threat to democracy, experts agreed in an online Associated Press article, is that AI deepfakes could erode the public’s trust in what they see and hear.
People won’t know whom to believe anymore and become more vulnerable to manipulation when choosing their leaders.
“A world in which everything is suspect — and so everyone gets to choose what they believe — is also a world that’s really challenging for a flourishing democracy,” noted one election information analyst.
Another expert said people “are busy with their lives, and they don’t have the time to check every piece of information. AI being weaponized in a political season, it could be extremely disruptive.”
A paper presented at the Davos economic forum regards AIpowered misinformation as the world’s biggest short-term threat.
As AI deepfakes become more sophisticated, it will be harder to stop them. “We’re not prepared for this,” said a cybersecurity expert. ”To me, the big leap forward is the audio and video capabilities that have emerged. When you can do that on a large scale and distribute it on social platforms, well, it’s going to have a major impact.”
The only way to stop AI from spitting out misinformation during elections would be to introduce a system of “quality control.”
Last April, the United States and the United Kingdom signed a memorandum of agreement to work together on evaluating the safety of advanced AI systems.
Under the agreement, the two countries will create AI safety institutes that will evaluate open and closed-source AI systems.
The European Union is also putting the finishing touches on its AI Act, which could be in force by this year.
The new law mandates that AI systems in the European bloc “should be overseen by people, rather than by automation, to prevent harmful outcomes.”
We need to set up the framework for AI governance as the world becomes more dependent on algorithms when making decisions.