National Unity Party (NUP) president LRay Villafuerte believes the Commission on Elections (Comelec) has a “good handle” now on checking campaign misinformation and disinformation on cyberspace in the runup to the 2025 elections, as a result of this electoral watchdog’s timely release of guidelines on the use by candidates and their supporters or of their rivals of artificial intelligence (AI) technology in next year’s midterm and Bangsamoro Autonomous Region in Muslim Mindanao (BARMM) polls.
“I believe our Comelec officials have a good handle now on monitoring and blocking digital disinformation and misinformation in the runup to the campaign season for the elections in May 2025, as an upshot of the Commission’s timely release of our country’s first-ever online campaign guidelines on the rational regulation of AI technology, with an eye to checking the use of deepfakes to unduly vilify or extol particular candidates,” Villafuerte said.
Although the use of AI technology in government processes can certainly improve the delivery of public services, especially with the automation of systems and data management, Villafuerte said, considering that, at this point, the potential for digital misinformation and disinformation appears boundless even as the capability of our poll watchdog and the rest of Government to respond to or counter it is limited.
As Comelec Chairman Greg Erwin Garcia himself and other experts have separately stressed in the past, Villafuerte said, “AI tools such as visual or image generators and voice-cloning software are readily available to shrewdly produce poll-related digital materials for or against candidates that can manipulate our electoral process during the campaign, balloting, vote-counting and post-tallying phases.”
“Hence, kudos to the Comelec, under the leadership of Chairman Greg (Garcia), in releasing this early the guidelines on AI regulation ahead of the campaign period for the 2025 elections, as AI tools like image generators and voice-cloning software are already being used in elections across the globe for producing and distributing AI-generated visual, audio or audio-visual presentations, and AI chatbots (robots) are being devised for poll bets to get in touch with voters,” he said.
Villafuerte said that, “With the way and speed that any information or image can so easily be spread through socmed (social media), I can’t imagine how pernicious its impact could be on digital campaigning in our upcoming elections, if the Comelec has opted not to regulate the use of AI and deepfake technology, given that Filipinos are among the world’s foremost users of socmed platforms such as Tiktok, Facebook, Messenger, Viber, WhatsApp, Telegram, Instagram and X.”
The poll watchdog last week released its Resolution No. 11064 covering the new guidelines on “the use of social media, artificial intelligence, and Internet technology, for digital election campaign and the prohibition and punishment of its misuse for disinformation and misinformation in connection with the 2025 national and local elections and the BARMM elections.”
“AI tools such as visual or image generators and voice-cloning software are readily available to shrewdly produce poll-related digital materials for or against candidates that can manipulate our electoral process during the campaign, balloting, vote-counting and post-tallying phases.”
The midterm balloting for national and local officials as well as for BARMM parliamentary officials will be held simultaneously on May 12, 2025.
The election period for next year’s polls shall be from Jan. 12 to June 11, 2025.
Promulgated by the Comelec, Resolution No. 11064 provides for, among others, the requirement for AI users to make a full disclosure of their use of such technology in their online election campaign materials; the taking down or blocking of poll-related information online that use AI tools but were not properly disclosed by their producers or distributors; and the penalties for the misuse of such materials as an election offense, pursuant to Section 261 of the Omnibus Election Code.
The new set of rules are limited to the regulation of the use, prohibition and punishment of the misuse of socmed, AI and Internet technology in the digital election campaigning for next year’s elections.
Approved for publication last Sept. 17, Resolution No. 11064 was posted on the Comelec’s website last week and certified by Consuelo Diola, Director IV at the Office of the Commission Secretary.
Under this resolution, the Comelec’s Task Force Kontra sa Katotohanan, Katapatan at Katarungan sa Halalan (Task Force KKK sa Halalan) is in-charge of implementing and enforcing the new guidelines, in cooperation with the Comelec’s accredited citizens’ arms and partner-organizations.
This Task Force is led by the Comelec’s Education and Information Department (EID) and its Law Department, with the assistance of the deputized law enforcement agencies and relevant government instrumentalities.
Earlier, in Minute Resolution No. 24-0486, the Comelec created Task Force KKK sa Halalan to combat the threats of disinformation and misinformation.
This panel used to be the Task Force Kontra Fake News, which was established in July 2023 under the Comelec’s Minute Resolution No. 23-0423.
“We hope that the new guidelines released by the Comelec would guarantee the rational regulation of poll-related audio, visual or audiovisual materials produced through AI technologies, including the prohibition of deepfakes designed to defame or falsely extoll particular candidates running for public office in the 2025 elections,” Villafuerte said.
A former governor who now represents CamSur in the House of Representatives, Villafuerte said: “The 19th Congress can support this Comelec plan to better prevent disinformation and misinformation during the 2025 campaign season by passing proposed legislation on AI protection.”
Villafuerte said the congressional approval of House Bill (HB) No. 10567, which he authored, is “timely and appropriate, given the Comelec’s issuance of AI-related guidelines and last month’s move by 50 organizations from the media, civil society and the academe to join the Commission in a new partnership to fight fake news ahead of the midterm polls in 2025.”
The Bicolano legislator was referring to “Panata Kontra Fake News,” which was launched at the GMA 7 headquarters in Quezon City last Aug. 30, on the occasion of National Press Freedom Day.
Weeks before the Comelec’s release of its digital campaign guidelines, Garcia told the House committee on appropriations during a public hearing on the poll watchdog’s 2025 budget proposal that the Comelec was about to issue its “historic resolution” or “guidelines on social media, guidelines on AI, guidelines on deepfakes including misinformation/disinformation.”
Before that, Garcia said he had recommended rules on AI regulation to the commission en banc because: “The abuse of AI technology and deepfakes undermines the integrity of elections and the credibility of public officials, candidates, and election management authorities.”
Garcia further said in a letter to the commission en banc, “The abuse of this technology in campaign materials, such as videos, audios or other media forms, may amount to fraudulent misrepresentation of candidates. This defeats the very purpose of a campaign, which is to fully and truthfully inform the voters about the elections and the candidates.”
“We hope that the new guidelines released by the Comelec would guarantee the rational regulation of poll-related audio, visual or audiovisual materials produced through AI technologies, including the prohibition of deepfakes.”
During a broadcast of the Malacañang Insider program, Information and Technology Secretary Ivan John Uy tackled the rise of an “scamdemic” using deepfake and generative AI tools, saying that these are “threatening the electoral process when unscrupulous individuals use them for political gain. Especially during elections, where they use them whether to malign or to discredit certain people by attributing quotations or phrases that are extremely unpopular and then making it appear that that person uttered those statements.”
Resolution No. 11064 defines:
* “deepfakes” as digitally manipulated images, videos or audio files “created using artificial intelligence to fabricate realistic representations of people, events or statements and falsely make it appear that an action, statement, or event transpired but did not occur in reality;
* “cheapfakes” as forms of “visual disinformation” for which authentic images or videos are re-contextualized to deliberately alter their meaning; and
* “soft fakes” as a type of media manipulation involving “subtle and often imperceptible alterations to content,” such as editing photos or videos, to influence perceptions or opinions without the more evident distortions characteristic of deepfakes.
Also, it defines “fake news” as misinformation, malinformation or disinformation deliberately presented as legitimate news and disseminated through quadmedia “with the intent to deceive, mislead or manipulate public opinion” or voter behavior; and “false amplifiers” as entities or mechanisms that “artificially increase the visibility or reach of digital partisan or political messages or narratives,” through the use of bots, fake accounts, or organized socmed networks.
“Hyperactive users” are defined by the Comelec resolution as “inauthentic individuals who engage in disproportionately high levels of activity on online platforms, including posting, sharing, or interacting with content at a significantly higher rate than the average user, with the intent to distort or manipulate online discourse or influence public perception using misinformation or disinformation.”
Under the Comelec guidelines, candidates or their representatives and registered political parties are required to register their poll-related socmed accounts, websites and Internet-based campaign platforms with the Commission’s EID within 30 days after the filing of the Certificates of Candidacy (COCs) until Dec. 13, 2024.
The guidelines require that all election propaganda and campaign materials of candidates and political parties that utilize AI technology “shall disclose the fact of its use and appropriate technology is employed to identify its authenticity and legitimate source.”
“Disclosures or disclaimers, which shall be clear, conspicuous, and not easily removed or altered, must remain visible or audible for a sufficient duration both before and after the campaign material is presented. The disclosure or disclaimer shall explicitly state that the content has been manipulated using artificial intelligence, providing a detailed explanation of the nature and extent of the manipulation,” it states.
Should the Comelec, through its Task Force KKK sa Halalan, detect the use of AI technology in any socmed-published or disseminated campaign materials without any proper disclosures or use of the required watermark technology for proper identification, these violators are required to explain why complaints should not be filed against them.
“A request for the removal, takedown, or blocking of the said content, social media or digital platforms, or accounts shall also be made by the Comelec, through the Task Force KKK sa Halalan to the technology platforms and providers, and/or the concerned law enforcement agency,” the guidelines states.
At the same time, the Comelec’s law department shall file election offense complaints against the violators, and the Commission shall request the technology platform or internet service provider, or file a complaint before the National Telecommunications Commission (NTC), Cybercrime Investigation and Coordinating Center (CICC), National Bureau of Investigation (NBI) Cybercrime Division, Movie and Television Review and Classification Board (MTRCB) and other relevant government agencies, to take down the subject content or the errant socmed media account, website or broadcast.
Villafuerte is seeking congressional action on his bill—HB 10567—to address fears about the lurking threats to the next polls posed by the proliferation of deepfakes or altered audio, video or audio-video presentations during the campaign period.
He said that HB 10567 aims to penalize producers or distributors for non-disclosure of “deepfakes, proposing cash penalties ranging from P2 million to P5 million for those who do not make any disclosure in these materials that these were altered with the use of AI or any other similar technology as well as on those who delete or tamper with such disclosures in these AI-modified voices or images.
A media report has identified OpenAI, MidJourney, DALL E2 and Dream Studio as among the softwares available online to anyone who wants to produce photo-realistic images with the use only of text prompts.
Villafuerte noted that among the most prominent victims of AI-generated misinformation or deception are incumbent American President Joseph Biden and his immediate predecessor Donald Trump, Ukrainian President Volodymyr Zelensky and even Pope Francis.
