MANILA, Philippines — The Commission on Elections (Comelec) and Department of Information and Communications Technology (DICT) should work together to take a “pro-active stand” against deepfakes, according to Rep. LRay Villafuerte.
Deepfakes are manipulated video and audio that mimic real people.
Villafuerte wants the Comelec and DICT to verify the authenticity of digitally altered photos or videos online of certain candidates in the run-up to the May 2025 midterm elections.
“With all of the candidates aspiring for the 18,000 elective posts, the Comelec and DICT need to work closely together on a system for AI (artificial intelligence) safety that can effectively monitor and take down deepfakes posted online between now and May 2025 to extol, vilify particular candidates,” Villafuerte, president of the National Unity Party, said.
He suggested that DICT, through its Cybercrime Investigation and Coordinating Center (CICC), should also put up an effective system of monitoring and taking down AI-generated deepfakes that are feared to proliferate online in the coming campaign season.
The use of AI and deepfakes during the campaign period will definitely have its consequences, the lawmaker from Camarines Sur said.
“Now that we know everyone who is running in the next elections in May 2025, the government needs to come up soon enough with an efficient way of monitoring and taking down illegal campaign materials posted online,” Villafuerte warned.
“(This is because the Comelec, DICT and CICC) themselves have publicly fretted about the likely spread of politically related AI content designed to spread misinformation and disinformation about specific poll bets,” he said.
At the recent launch of Cyber Awareness Month, DICT Undersecretary Jeffrey Ian Dy warned the public about a surge in AI-generated content in the election season, as his office has monitored the rise of AI content that are either misleading or spreading misinformation and disinformation.
“What sets this election apart is that this is going to be an election that is heavily digitized. And I think our regulations have not yet adopted to that,” Dy said in a report.
“This is now a major problem that we never experienced in past elections, like the use of disinformation through AI-generated content,” he said.
Dy said that, aside from deepfakes, another type of AI-generated content being used to spread false information is “vishing,” or calling by copying the voices of other persons.
He claimed that three major social media giants – Meta, Google (YouTube) and TikTok – have committed to assist the government whenever it detects AI-generated misleading content.
“What we actually want the social media network to do is to label them: ‘This is misleading,’ ‘this is AI-generated,’ ‘this is taken out of context,’ ‘the source content can’t be verified.’ Those kinds of labels would help our people in assessing (fake content),” Dy said.
Meanwhile, the Comelec has named 156 organizations and political groups set to join next year’s midterm elections.
Comelec, in a resolution, announced that the 156 organizations and political groups shall first participate in a raffle to determine their order of listing on the official ballot. This will be held on Oct. 18 at 9 a.m. — Mayen Jaymalin