Will AI turn the tide in midterm elections?

By Kiran N. Kumar

Indian American Congressman Ro Khanna has strongly pitched for regulating political advertising on social media

With the midterm elections just a week away, several artificial intelligence (AI) apps have been making rapid inroads into the electioneering with a profound belief that it is possible to use technology to steer and predict election results.

CivixAI, for instance, successfully tested its AI algorithm in Florida’s Democratic primary election in August. Its findings prior to the elections projected that Democratic voters favored Tallahassee Mayor, Andrew Gillum over all other candidates.

Similarly, for the Republican primary, it found Ron DeSantis stronger than Florida Commissioner of Agriculture, Adam Putnam. Accordingly, both Gillum and DeSantis won the respective primaries.

Upbeat over the accurate prediction, Ryan Gragg, Director of Engagement and Co-Founder of CivixAI says that the technology could be used in the upcoming US midterm elections on Nov 8.

READ: Why Artificial Intelligence remains a distant dream? (June 16, 2022)

“The specific problem we’re solving is to understand and predict, in advance, what action voters and citizens will take in a particular scenario based on their cognitive bias. For instance, we may want to predict who voters will elect in a specific race on Nov 6.”

This is not the first time AI was used to predict election results as in the 2016 general election, Indian startup MogIA was able to accurately predict Trump’s win based on data analysis gleaned from social media.

Historically, midterm elections generate lower voter turnout than presidential elections with a 40% turnout but nonetheless the outcome serves as a referendum on the sitting president’s performance.

Otherwise, midterm elections are being held for all the 435 seats in the House of Representatives, and 34 of the 100 seats in the Senate, besides 36 of the 50 state governors.

But the AI technologies are often overwhelmed by social media bots deployed as in the 2016 presidential election to influence the poll in favor of Donald Trump while the same bots countered his campaign at every point and worked against him in the 2020 election.

Whether presidential or midterm, elections rely heavily on social media platforms and bots. Emilio Ferrara, a researcher at the USC Information Sciences Institute, had warned in 2016 that the robots can overtake political debates to decide the outcome.

READ: Smart homes on rise, will Artificial Intelligence take over next? (May 14, 2022)

“Software robots (bots) masquerading as humans are influencing the political discourse on social media as never before and could threaten the very integrity of the 2016 US Presidential elections.”

One estimate puts the figure at about 400,000 bots which were engaged in the political discussion about the presidential election, sending out roughly 3.8 million tweets, or cornering about one-fifth of the entire political debate.The bots were able to do so by stealing online pictures, giving fictitious names, and cloning biographical information from existing accounts.

So sophisticated these bots were that they could “tweet, retweet, share content, comment on posts, while clicking on ‘like’ for candidates, enhancing their social influence by following their accounts and even engage in human-like conversations.”

While Twitter found that more than 50,000 Russia-linked bots were active on its network during the 2016 election, reaching more than 1.4 million Americans, a study by Yale University in 2020 said that these accounts were automated as the code detects tweets and retweets them automatically, running 24 hours a day, far more active than a human being.

READ: Canada’s defense scientists develop a model to enable trust in Artificial Intelligence and Autonomy(January 28, 2021)

Certain kinds of content were amplified using bots to make it look like there are more people talking about it. Beyond bots, trolling and disinformation remain permanent challenges during elections.

“One thing that gets lost is disinformation and online media manipulation in 2020 does not look the way it looked in 2008 and even 2016,” says Zarine Kharazian, an assistant editor at the Atlantic Council’s Digital Forensics Lab.

As trolls have become the norm since the run-up to the 2020 election, cybersecurity experts warn that the bots are playing a crucial role amplifying them.

“Overwhelmingly, we see much of the disinformation, vastly more of the disinformation, is produced by US sources,” said Timothy Frye, a Columbia University political science professor. “And what we see Russian sources doing is amplifying them.”

Stressing that it’s the major responsibility of social media giants to be more vigilant on social media, Indian American Congressman Ro Khanna has strongly pitched for regulating political advertising.

“There is a need for a regulatory body that looks into these advertisements and more importantly notifies if there is false political advertising,” said Khanna reiterating the need for social media companies to shoulder the responsibility to prevent viral false news, particularly at a time when the tone of politics in the country has become very hostile.

Now that Elon Musk has taken over Twitter despite the perceived presence of millions of bots, the writing on the wall is clear — Own technology first, politics follows next.