Perilous AI misuse

The unregulated use of artificial intelligence (AI) could significantly disrupt elections in this country, including the midterms in May and the even more critical one to follow in 2028, exacerbating existing misinformation, voter manipulation, and political instability.

As candidates for elective positions in the May elections push their campaigns in the last stretch, voters should ask themselves if the audiovisual clips and other content they see on social media are facts-based authentic, or merely AI-generated.

AI power bots and fake accounts can amplify propaganda, making certain candidates appear more desirable than they really are to voters. Political bets using such tools may resort to personalized disinformation, with AI chatbots made to micro-target voters with tailored lies.

A case in point was the use of deepfakes and AI-generated propaganda allegedly by Indonesia’s now President Prabowo Subianto who had once been perceived as a hardline military figure and, through AI-generated imagery, rebranded himself as a “cuddly grandpa.”

His campaign posted AI cartoons and TikTok videos that softened his image, appealing to Gen-Z voters. AI reshaped voters’ perceptions, particularly the younger Indonesian electorate, which contributed to Prabowo’s landslide win.

Unregulated AI on such platforms as TikTok, Facebook, and YouTube could push divisive content, deepening polarization among the electorate, even as fake AI-generated influencers “astroturf,” or simulate fake public support for certain candidates.

To recall, then candidate, now India’s Prime Minister Narendra Modi’s campaign team in the 2024 elections used AI voice cloning to deliver personalized messages by Modi in different dialects, e.g., his addressing voters in South India where Hindi isn’t dominant, thus improving outreach.

Another trickery was the AI-generated “interview” of deceased politician, M. Karunanidhi, which was circulated in an effort to influence Tamil voters. There, too, were AI voice clones used to spread such false claims as Muslim voters getting double ballots, thus inciting religious tension. In all, AI deepened political polarization and allowed campaigns to bypass traditional media, manipulating voter sentiment.

Such instances of fakery should be borne in mind particularly with the coming midterm elections seen as an extremely high-stakes political exercise for pro-China bets. There’s much talk about the existence of so-called Chinese-controlled troll farms using AI to craft and spread pro and/or anti-candidate narratives to favor Beijing’s candidates in the Philippine elections.

AI could likewise automate cyberattacks on election data bases and government websites, causing distrust in the election results.

Scammers, as in the case of the last elections in India, could clone voices to intimidate voters, impersonating officials and harassing voters with fake threats, e.g., “you’re registered illegally, don’t vote!”

This is also the season for surveys of all sorts, whether for the benefit or not of candidates. In all, people should be vigilant about the objectives of those circulating false information as well as spreading untrue election results to undermine the people’s trust in media and the democratic process altogether, creating chaos for their nefarious ends.

Facebook honcho Mark Zuckerberg, in an attempt to suck up to Donald Trump, may have scrapped fact-checking on his social media platforms, but local fact-checkers can still do their all too vital job here, highlighting media literacy efforts to mitigate whatever damage may be wrought by fake news with the use of AI.

The last thing we would like to see as a result of unregulated AI is a body politic saturated with fake news, mass disinformation resulting in bitterly contested results, violent clashes and a legitimacy crisis.

How can government protect the elections against AI-using brigands who hope to shape the election toward their own ends?

Well, it is never too late to ban AI-generated deepfakes, mandate transparency for AI-generated political content, and strengthen cybersecurity to prevent AI-powered hacking of election systems.

Efforts must likewise be expended by stakeholders to push for public AI literacy to help the electorate spot fake content and regulate social media algorithms to prevent AI-driven echo chambers.

These are essential measures to safeguard the legitimacy of the midterm elections in May, and the even more crucial political exercise to come in 2028.

Not having the necessary security protocols in place would have dire consequences, with unregulated AI wreaking havoc, turning Philippine elections into a disinformation battleground and, God forbid, in its wake, democracy in the country a precious casualty.