Put AI to good use in midterm polls, data experts tell Comelec

THE COMMISSION on Elections (Comelec) should use artificial intelligence (AI) to improve the integrity of the midterm elections next year amid worries about potential disruptions, analysts said.

Instead of trying to police the use of AI on social media, the agency should establish clear guidelines for candidates and supporters, Sam V. Jacoba, founding president of the National Association of Data Protection Officers, said in a Viber message last week.

The rules should include transparency requirements, data protection and guardrails on AI’s influence on voter behavior, he said.

“AI is a tool that the Comelec can use to make our elections clean, transparent and efficient,” he said. “Banning its use for elections will stifle innovation and will not be the best way to harness its potential for good governance.”

He said AI could be used to detect if a candidate is ineligible or clean the voter database.

Comelec Chairman George Erwin M. Garcia said they don’t plan to ban AI in next year’s elections.

“It was exaggerated to the point that I was asking for a ban on AI, although I honestly believe and the Commission En Banc believes that that cannot be done,” he told BusinessWorld by telephone at the weekend.

He said there are certain aspects of AI that are “so good” and certain facets that are evil. He added that the Comelec would issue the rules on AI and ethics by August.

He said deepfakes should be prohibited because it misrepresents the facts. “The problem with this is that social media will be included [in the potential ban]” he said. “That’s a very touchy issue because there may be a very thin line between constitutionality and unconstitutionality.”

Mr. Garcia said the Comelec would partner with social media platforms after the release of the guidelines.

“We are requiring the candidates to submit social media accounts that they are going to use,” he said. The election body could then ask social media platforms to remove unauthenticated accounts that promote a candidate.

Mr. Garcia said they might not have enough manpower to monitor social media and AI use during the campaign for local and national positions. They might have to seek the help of citizen’s and interest groups, stakeholders and IT experts.

The Comelec has kept the Pastoral Council for Responsible Voting and the National Movement for Free Elections as citizen’s arms for the 2025 elections.

On July 20, the election body introduced a task force that would monitor fake information against the Comelec, its officials and the election process.

The Philippines has yet to pass a policy similar to the European Union’s AI Act, though the Trade department has launched a national AI roadmap.

Ronald B. Gustilo, national campaigner at Digital Pinoys, said not all AI-generated materials are bad.

“The Comelec should let candidates and supporters use AI-generated materials, as long as these are law-abiding and comply with existing rules and regulations,” he said in a Viber message. He agreed that deepfake videos meant to malign candidates should be banned.

Allan S. Cabanlong, regional director for Southeast Asia at Global Forum on Cyber Expertise, said it would be difficult to trace the origin of deepfake videos.

“There is a corresponding penalty in the Cybercrime law, although there is no cybercrime definition,” he said by telephone. Still, the law covers deepfakes since these involve altering data, he added.