By Ruth Cabal

The issue surrounding Senator Ronald “Bato” dela Rosa’s post featuring a video made with artificial intelligence (AI) exposes how many Filipinos seem to be still grappling with the concept of AI.

On June 13, he shared on Facebook videos of two students being interviewed. They said they are not in favor of the impeachment of Vice President Sara Duterte, questioning its motive and describing it as selective justice and politically-motivated. Dela Rosa captioned it with “Buti pa ang mga bata nakakaintindi sa mga pangyayari. Makinig kayo mga yellow at mga komunista!”

(“Good for the youth, they understand what’s happening. You better listen, yellows and communists!”)

In the comments section, many agreed with the senator. Katherine Aquino and Evelyn Dupyawon wrote the same remark, “Very well said.”

But others called out dela Rosa, noting the video was made with AI. “AI generated meansfake,” commented Ariesan Dencaldeo. Jake dela Cruz suggested, “Be digital literate.”

The video shared by dela Rosa was originally from Ay Grabe’s Facebook account. It was not labelled as AI-generated, though the caption included the hashtag #AI. On the lower right of the video, there’s a small, barely visible watermark of Veo. Many might not be familiar with it as Google’s AI-powered filmmaking tool. Google says on its website that Veo can generate realistic videos based on prompts.

Doc Ligot, an AI expert and Founder of Data and AI Ethics PH, says there are ‘context clues’ or hints which may show the video was AI-generated. He points out the outdoor scene is obviously not in the Philippines. What may seem like a tricycle in the background is not the kind commonly seen here. Even the signages, mostly unreadable, don’t look like they are from buildings in the Philippines.

Ligot explains, “It was AI that generated the scene. Maybe in the data of AI, a typical Thai or a typical Filipino look very similar – which may be true.“

Facebook has since labelled Ay Grabe’s post as false information, according to its third-party fact checkers.

Ligot says AI-generated content is now changing the way we validate information. If in the past, seeing is believing, it seems to be the opposite now. Before, simply taking a picture or video of an event is enough to prove its originality and authenticity.

“Video can’t be trusted anymore, audio can’t be trusted anymore. So parang guilty before proven innocent… Eventually it will come to a point where you can’t really tell (whether it’s AI or not.) Lalo na if the person generating the media is quite savvy also with editing and how media works. They can make it very flawless.”

WHAT IS AI?

In the past, the phrase ‘artificial intelligence’ would usually conjure images of a robot or some sort of sci-fi setting. Ironically, many of AI’s uses now are aimed at imitating present reality. While its definition includes automation and computers, the human aspect can never be disregarded.

Britannica Online defines AI as a term “frequently applied to the project of developing systems endowed with the intellectual processes characteristic of humans, such as the ability to reason, discover meaning, generalize, or learn from past experience.”

A University of the Philippines paper on AI principles declares, “AI systems embody computational structures that mimic human or animal cognition to process data, learn from experiences, and decide, plan, and act autonomously to satisfy a programmed objective.”

Simply put, it’s artificial intelligence as opposed to human intelligence. After all, it’s still humans who create AI tools and tell these systems what to do.

In the case of the video shared by Senator dela Rosa, this is called a deepfake, according to Ligot.

A deepfake, he explained, is a combination of deep learning and fake media. “Deep learning is a subset of machine learning, the science of AI. Essentially it’s any process which looks at data and creates predictions or new data. So I’ll get photos of you so I can create fake photos of you, because it learned your facial features and patterns.”

Ligot adds, the process of creating deepfakes has become easier now with many text-to-media tools available online such as Veo, Kling, and Sora, among others.

“Now you can just tell the tool, give me a scene… And separately you can clone voices. Then I can get that voice to say anything I want. So combining the two, yun na, you have synthetic video and audio.”

AI-GENERATED CONTENT AS FAKE NEWS

Dela Rosa later deleted the video he shared on Facebook and posted, “I don’t care if this post is AI-generated or Manobo-generated because I am not after the messenger, I am after the message.”

Palace Press Officer Claire Castro countered, saying dela Rosa’s message was disinformation or fake news. “Nakakaduda, mas nakakawala ng tiwala kung mismong sa matataas na opisyal nanggagaling ang mga disinformation at fake news. Responsibilidad nila ‘to. Dahil ang bawat salita na binibitawan nila sa taumbayan, lider sila, yaan ay totoo sa pandinig at pananaw ng bawat isa.”

(“Suspicious, and all the more erodes trust when it’s a top government official who’s the source of disinformation and fake news. It’s their responsibility. Because every word they say, as leaders, is taken as the truth in the ears and eyes of the public.”)

Danilo Arao, associate professor of the Department of Journalism at the University of the Philippines, says Dela Rosa should be responsible enough as an elected official to ensure his statements are factual – whether in person or online.

“Apparently Senator Dela Rosa does not know much, not just about AI but even the basics of social media consumption. Hindi mo pwede paghiwalayin ang mensahe sa pagiging tama ng mga datos… Mahihirapan tayo sa pagbuo ng opinyong pampubliko.”

(“You cannot separate the message from the accuracy of data. It will be hard to shape public opinion.” )

The issue comes as a new study shows more Filipinos are turning to AI-powered tools for news and verification of information. According to the 2025 Reuters Institute Digital News Report (DNR), 9% of adult Filipinos surveyed say they get their news from AI chatbots. 10% use bots for verification. ChatGPT was the most used chatbot, followed by Meta AI and Google Gemini.

Overall, online sources were still the top news source of Filipinos (83%), followed by social media (66%) and television (46%).

HOW TO SPOT AI-GENERATED CONTENT AND DEEPFAKES

Detecting AI is difficult, says Ligot. While there are already some AI tools that claim to detect AI, these may not always be accurate. Also, such tools may only exist for some time because technology will keep on improving.

He advised going back to the basic process of verifying whatever information is generated by AI. Be on the lookout for certain themes that are highly susceptible to fake content such as politics and reputation, commerce and potential scams.

When viewing AI-generated content, one should also figure out its objective. For example, even before the advent of AI, satire or propaganda, meant to ridicule, criticize or promote something are viewed differently from political news. The same is true if an AI-generated video is meant for entertainment, for example, or simply just to illustrate a personal story. For these, a viewer’s standards of accuracy and truthfulness may not be as strict, depending on what the AI-generated content is trying to achieve.

“There’s no data set of truth for AI. Everything is debatable… The harder part is being critical. We have to automatically train ourselves to be skeptical of anything we see, lalo na online,” says Ligot.

He adds what could make it easier and more ethical for users, and social media and other online platforms is labelling content as created or modified using AI. Facebook, for example, has an AI label button. Its info page says this “helps establish transparency and trust.” Facebook requires AI info labels for photorealistic video or realistic-sounding audio. The platform says it may also automatically label AI content, as determined by its system.

For Arao, media education and literacy is the long-term solution that can help detect not just AI-generated content, but all other forms of fake news and disinformation. He proposes this to be taught at K-12 and tertiary levels, especially since this age group is among the most active consumers of online and social media content. Arao stresses it’s not enough that they have to contend with AI and fake news but more so, with its implications out there.

“The ‘real world’ consists of truth-seekers who get vilified and liars get exalted, and the youth should be equipped with the necessary knowledge to distinguish between the truth and the lies.”