When veteran journalist Ruth Cabal became the victim of artificial intelligence (AI) manipulation, there was a very brief moment that the news anchor — who has read thousands of news items on camera throughout her career — actually thought it was her speaking in the video.
She said there were instances in the past when her image was used in “obviously fake” Facebook games, but in that one particular video that gained online traction in December last year — where her likeness was seen promoting a get-rick-quick scheme — she was caught off-guard.
“The tone really sounded like my voice,” she said. “But when I played it further, when the word ‘pesos’ was pronounced, I thought, that’s not how I would pronounced the word ‘pesos.’ That’s when I came to the conclusion it was fake.”
Cabal recalled how she immediately reported it to her employer, who told her she was already the second anchor in the newsroom to have been victimized by deepfakes, which are realistic but fake media created using artificial intelligence.
“It was a very scary experience. Because what if one family invested their savings, because they thought they saw it in the news, reported by Ruth Cabal?” she lamented.
Deepfakes featuring journalists in the Philippines have become increasingly common in the past year, and the threat of AI-related disinformation grows larger as the country heads to another general election year in 2025. It’s a conversation that took center stage when Rappler held its Social Good Summit (SGS) on October 19.
In the United States, over a dozen states have approved legislation to regulate the use of deepfakes in the run-up to the November 2024 presidential election, although federal measures continue to languish in the US House of Representatives.
Here at home, chances of a measure that would combat disinformation before next year’s midterm vote being passed are close to zero.
Internet propaganda
The Philippines is no stranger to internet propaganda that sows confusion and misinforms.
One study shows that widespread “fake news” impacted the 2016 polls, which catapulted Rodrigo Duterte to the presidency, and the 2019 midterms, which solidified his grip on power.
In 2022, President Ferdinand Marcos Jr.’s landslide election victory was fueled by “disinformation, coordinated amplification, and use of an extensive network of anonymously-managed pages,” based on Rappler’s investigation.
During that time, watchdog Movement Against Disinformation observed the lack of urgency on the part of social media platforms in responding to incident reports.
“For example, there was one candidate who was promising to distribute upon election hordes of hordes of gold. What we did, we went to the major social media [platforms], and we complained. It took them a long time to get back to use. And that was experienced by even the government regulators. That is a legal environment that we are working under,” MAD Chairman Rico Domingo also told the SGS audience at the De La Salle University in Manila.
Nearly two years since, Cabal’s case highlights an underdeveloped mechanism to expedite the takedown of false content online.
“As a journalist and as a victim, there are so many avenues. Where do I go? Philippine National Police Cybercrime, National Bureau of Investigation, Department of Information and Communications Technology (DICT), then you also mentioned a task force. I don’t really know. I don’t know if that’s a problem,” Cabral said.
According to the DICT, people can report incidents to them, but the process isn’t perfect.
“There’s now Executive Order 58. Email us at ph-cert@dict.gov.ph,” DICT Undersecretary Jeffrey Ian Dy replied. “But we have to be honest. Let’s face it, we cannot report for 125 million Filipinos. NCERT [National Computer Emergency Response Team] focuses on critical information infrastructures and high-risk individuals, including media personalities.”
Short-term solutions
Even the country’s highest leader became a purported subject of deepfakes earlier this year.
In April, Malacañang warned the public against a deepfake audio of President Ferdinand Marcos Jr. ordering a military attack. In July, a video flagged as suspicious by deepfake detection tools showed a person — which resembles a younger chief executive — taking drugs.
Subsequently, the Commission on Elections launched Katotohanan, Katapatan at Katarungan (KKK) sa Halalan, a task force that seeks to counter AI-driven disinformation on television, radio, print, and online.
The poll body, more importantly, has released guidelines detailing regulations on the use of AI in campaigns.
The Comelec’s resolution requires candidates to disclose if their campaign materials use AI, and prohibits them from spreading disinformation through deepfakes, “fake news” content, fake accounts, bots, and coordinated inauthentic behavior, among others. Violators may be the subject of an election offense case.
The Comelec envisions that collaboration with social media platforms would result in a quick removal of false and misleading content online, posted by candidates’ accounts that are registered with the poll body.
But even Chairman George Garcia is aware that this is a band-aid solution.
“We can regulate the candidates, and they will be afraid. We will threaten them with disqualification, election offense cases, and one to six-year imprisonment. But how about the supporters? Candidates always claim supporters are the culprits, not them,” he said.
The reality is that the Comelec, as an implementer of laws, is working with an antiquated election code, a measure that precedes the Philippines’ post-EDSA democracy. It is one that does not have provisions on campaigning on social media, which did not exist at the time of the bill’s passage in 1985.
The only legal justification for the poll body to come up with social media guidelines for campaigning is a throwaway provision from Republic Act No. 9006 or the Fair Election Act of 2001, which states that election propaganda in “any other medium” must be subject to Comelec supervision.
“At this point, the Comelec promulgated the rules in accordance with RA 9006. But I will readily admit to all of you, as regards its legality and constitutionality, it might be questioned later on,” he said.
Legislation
With an inadequate legal framework for the online space, the Comelec, the DICT, and even anti-disinformation advocates see the need for the passage of a social media regulation bill.
“Without a social media regulation law, properly crafted, we don’t have hope here,” Domingo said.
When Rappler CEO and Nobel Peace Prize laureate Maria Ressa faced the Senate in 2022, she said the key to combating disinformation is crafting a law that would make social media platforms liable for allowing lies to spread online.
The push is not without its risks. If the bill goes to the wrong hands at the legislature — to those with little sensitivity to freedom of expression — the law could end up suppressing speech of individual users, instead of taking big tech to account.
A number of bills have been filed in Congress seeking to regulate social media, but none has gained significant traction. The measure is not among the priority bills of the incumbent administration.
The DICT, however, said it is working on an executive version of the bill, and promised to put a premium on the right to free speech.
“Don’t worry. We’re not in favor of draconian measures like takedown of posts. We just want everybody to be aware. We’re in favor of tagging. Progressive violations can mean taking down of the posts, but we want that space where you can still be more responsible and say, ‘oh, don’t worry, this is just political satire.’ We just tag it, that simple. This is fake. It’s just for fun. We just want you to dance along with BBM, you know?” Undersecretary Dy said.
“We are promoting and actively engaged with the legislature on the passing of a social media regulation bill. I think this is now ripe and mature,” he added.
– Rappler.com