A video that has gone viral on social media shows resigned Ako Bicol party-list representative Zaldy Co allegedly admitting that the Dutertes have ordered him to implicate President Ferdinand Marcos Jr. in the budget insertion and flood control corruption scandal. This is fake.
At least two Facebook pages and one TikTok user published the fake video on Nov. 14. One of those bore the following caption:
“Ito na Zaldy Co nagsabi na ng Totoo mga Duterte pala ang salarin (Here it is. Zaldy Co said the truth. The Dutertes are the culprit).”
The 17-second video showed Co saying the following:
“Inaamin ko sa publiko na ako ay napag-utusan lang din lamang ng mga Duterte para siraan ang ating pangulo. Umaamin ako na lahat ng binasa ko ay galing mismo sa nagbigay sa akin at ako ay napag-utusan lamang.
(I admit to the public that I was only ordered by the Dutertes to defame the president. I also admit that everything I have read came from somebody who gave it to me and I was just told what to do).”
Co never said such a thing. Artificial intelligence was used to create the spurious video.
From Nov. 14 to Nov. 16, Co published a three-part recorded video accusing Marcos and resigned House speaker Martin Romualdez as the culprits behind the P100-billion worth of insertions in the 2025 national budget, from which they supposedly received P56 billion in kickbacks.
The former lawmaker then published the full, unedited nine-minute video on Nov. 17. Nowhere did Co ever say that he was coerced by the Dutertes to paint Marcos in a bad light during the full runtime of the video.
One key observation between the original and fake video is that Co is seen looking directly into the camera for 17 seconds straight in the AI-generated version. In the original, the former lawmaker spent most of his talking staring down to read his script.
VERA Files Fact Check reached out to the Deepfakes Analysis Unit (DAU) of the India-based Trusted Information Alliance to further analyze the video. “[Zaldy Co’s] teeth are all jumbled together into an off-white patch, instead of being distinct,” DAU said.
“Towards the end of this video, his mouth remains open for a split second where no teeth are visible,” they added. “Overall, teeth, lips, and mouth regions are usually the best clues to determine the authenticity of videos. Aspects like body language, and details of the clothing worn by subjects is beginning to become relevant to investigations too.”
DAU ran the spurious video through Hive AI audio classifier and deepfake classifier, Hiya Audio Intelligence, Aurigin.ai and Eleven Labs Speech Classifier. Aurigin.ai, a Swiss deep-tech company with an advanced audio deepfake detection engine, and Hive AI’s deepfake classifier found signs of AI tampering in the video.
VERA Files Fact Check has debunked a different AI-generated content of Co earlier this month. (Read FACT CHECK: Videos of Zaldy Co spotted by netizens AI-GENERATED)
TikTok user heymangboyy first published the AI-generated video, receiving over 14,800 reactions,7,800 comments, 3,800 shares and 594,000 views. It was then reposted by FB pages NARDZ TV51 (created on March 12, 2024) and Lokong Rekha (April 27, 2023), collectively garnering over 8,800 reactions, 4,700 comments, 3,100 shares and 1 million views.
