The Commission on Elections led by chairman George Garcia along with the Manila government remove illegally placed campaign materials along Oraquieta cor. Recto Avenue on October 20, 2023, a day after the start of the campaign period for the Barangay and Sangguniang Kabataan elections. Maria Tan, ABS-CBN News/File 

MANILA (UPDATED) — Commission on Elections (Comelec) chairman George Erwin Garcia has formally sought to “prohibit the use of deepfake and artificial intelligence" in campaign materials for the 2025 polls.

• Watchdog urges creation of 'AI code of conduct' ahead of 2025 elections

In his May 28 letter to the Commission en Banc, Garcia underlined the serious effect of such technology during elections.

“A growing concern among several election management experts in various jurisdictions is the threat of Artificial Intelligence (AI) technology and 'deepfake' to sow deceit and misinformation between and among voters and candidates for elections,” Garcia said in his letter.

“The abuse of AI technology and 'deepfake' videos undermines the integrity of elections and the credibility of public officials, candidates and election management authorities,” he added.

• Are AI-enhanced photos dishonest? Senator seeks Comelec rules

Interviewed minutes before the Commission en banc session Wednesday, Garcia said that candidates must be banned from using AI and deepfake technology during the campaign period.

“Kukumbinsihin ko po ang ating Commission en Banc na ipagbawal namin sa campaign ang paggamit ng deepfakes at AI sapagka't madami po ang nalilito, misrepresentation, madaming hindi makatotohanan na nangyayari kapag ginagamit po yan,” Garcia said.

(I will convince the Commission en Banc to ban deepfake and AI use during the campaign period because it promotes confusion, misrepresentation, falsehood.)

• Philippines must be part of the AI revolution, Marcos says

DICT ready to assist

The Department of Information and Communications Technology said it is ready to assist the Comelec in keeping deepfakes from influencing the elections.

DICT Undersecretary Jeffrey Ian Dy told ABS-CBN News that while the agency has yet to be informed of the request, it is ready to help the Comelec in implementing this “bold move.”

“Ang DICT is the chairman of the Comelec Advisory Council. We’ll be more than happy to give them assistance if they would require it,” Dy said.

“I would say that it’s a bold move for the Comelec and we welcome that particular move,” he added.

Dy admitted that while AI creations have become more sophisticated and realistic making it more challenging to identify it from a video taken by a human, the measure may hinder certain actors from abusing the technology in their campaign materials.

As a first step, Meta’s feature labeling AI content can be utilized for transparency according to Dy.

“It could be that somebody would have copied a candidate saying that he or she is withdrawing from the campaign when in fact, that is not true,” Dy said. 

“(Sa Facebook), I think It now asks you to tag if what you’re uploading is already AI. I think that’s a good move towards transparency. And I think that if you haven’t been transparent about it, for example you know it’s AI for a fact but you didn’t tag its AI, that could be the first part of implementing this policy of Comelec,” Dy explained.

Garcia’s memorandum cited the unauthorized use of the voice of President Ferdinand Marcos Jr. in a manipulated video to deceive viewers as one of the reasons for the request to prohibit deepfakes in campaign materials.

Doubts and possibilities

But former congressman Edgar Erice questioned how such regulation could be implemented.

“Pwedeng regulation, pero papaano mo ire-regulate? Eh yung kandidato, may lumabas na ganun, sabihin lang nya, ‘hindi naman galing sa akin yan.’ Parang ano lang yan, parang poster, bawal dito, pero naglalagay pa rin. Sasabihin nung kandidato, eh hindi naman ako yung naglagay nyan eh, yung kalaban ko,” Erice said.

(Regulation is possible. But how can you regulate? A candidate can easily say that ‘that did not come from me.' Like when putting up campaign posters. The candidate can easily say that his political opponent was the one who posted his campaign posters.)

Cybersecurity expert Oscar Visaya advised the government to invest in infrastructure for its security architecture.

Aside from deepfakes, Visaya said the government should be prepared to guard the elections from other cyber threats.

“When the elections are questionable, the country becomes destabilized… In terms of systems - Comelec applications and systems - they need to be protected at all times,” Visaya said.

Visaya suggested the use of “good AI” to counter threats from bad actors also using AI for cybercrime.

“Even if access has been granted, it needs continuous monitoring so that it’s not something that they’ve stolen the credentials of a valid user. In that case, we’ll be using AI as well to defend ourselves so that we can see that it’s not real. The user who has been regularly using this at these times of the day, but now at a different time of the day - we can detect these anomalies with AI,” Visaya said.

Meanwhile, Art Samaniego of Scam Watch Philippines can see the feasibility of banning AI and deepfakes in next year’s campaign activities.

“Dapat siguro pagbawalan yung source. Ano ngayon yung mga source ng AI-generated content? Social media, Facebook, X saka Instagram,” Samaniego said.

(We should ban the source. And what are the sources of AI-generated content?)

 • OpenAI releases guidelines to gauge 'catastrophic risks' of AI

But for Senator Win Gatchalian, a clear policy is needed to ensure the effectiveness of Garcia’s plan to ban AI use during the campaign.

“We have to clarify din kung ano yung iba-ban sa AI. Dahil masyadong malawak yung application ng AI eh. May mga applications na legal, pero may mga applications din na illegal,” Gatchalian said.  

(We have to clarify what to ban in AI because that is too vast. There are legal and illegal applications of this technology.)

RELATED VIDEO