Deepfake videos are on the rise, making the government and experts concerned about the spread of disinformation through this medium.
As seen in Raffy Tima’s report on “24 Oras,” Thursday, deepfakes are generated through artificial intelligence (AI) and they can copy a person’s likeness, make them speak in other languages, or deliver a false message to viewers.
According to Carlo Nazareno, the Director for Rights of Democracy.net.ph, producing deepfakes are now easily accessible to the public.
“Gagamit ka lang ng app, makakapag-deepfake ka na. Tsaka kailangan mo lang ngayon ng gaming PC, kaya mo na gumawa ng Deepfake as sarili mong computer,” he said.
What’s even more alarming is the highly realistic rendering of images, which can make people easily believe in deepfakes.
“Very realistic na siya ngayon in certain respects lalo na kapag ‘yung gumawa ng deepfake is marunong gumamit ng visual effects or visual effects artist,” Nazareno said.
However, there are still ways to distinguish a deepfake video from a real one.
“Tingnan mo na mabuti ;yung ngipin habang nagsasalita. Kasi, minsan nagwa-warp ‘yung ngipin, pati ‘yung labi. So, ‘yun ‘yung isang bantayan. Pero ‘yung pinakababantayan natin ‘yung mensahe mismo. Tanungin natin kung may parang agenda ba itong video na pinapanood natin,” he said.
Dominic Ligot, a data analyst and AI ethics advocate, seconded that scrutinizing the message of a video is a good step towards identifying its veracity.
“Dapat matuto tayo to think of the context, ano ba ibig sabihin nun? Ano ‘yung intention or mukhang intention nung content or video , may sinisiraan ba? Or meron bang hina-hard sell na something products or services. Usually ‘yun ‘yung dalawa e, reputation attack and hard selling,” he said.
Oftentimes, deepfakes target the emotions of social media users.
“Gusto ka nilang ginagalit, pinapalungkot ka para ‘yung gagawin mo sa daliri mo either to share or mag-react ka doon sa post. E ‘yun kasi translates into money for the publishers,” said Celine Samson, Head of Online Verification of Vera Files, the third-party fact-checking partner of Facebook owner Meta.
Aside from the message, Ligot also said that the source of the video is another factor in determining whether it is fake or not.
“One of the best places to start is sino ‘yung source? Sino nagsabi niyan, is it a credible source? Or some random website lang, some random page, dun ka magsisimula and then later, ano ba ‘yung ina allege ng post na ‘yan,” he said.
Although there are ways to check for deepfakes, Janina Santos of Democracy.net.ph said it is still important to deliver the truth ahead of the disinformation.
“Importante po ‘yung tinatawag din na prebunking kung saan sinisigurado natin na mauuna tayo sa pagpapakalat ng totoong impormasyon,” she said.
“Gusto po natin na mauna para po ma-frame po natin ‘yung ating mga kababayan na ito talagang ‘yung totoong nangyayari. We can call it truth inoculation.”
Department of Information and Communications Technology (DICT) Undersecretary Jeffrey Ian Dy said the ASEAN region still has no laws covering online harms, so everyone must remain vigilant with information they see online.
“Wala rin tayong regulatory powers over social media networks and over OTBs, ‘yung mga chats, so that actually emboldens them to do things kasi hindi naman sila nare-regulate, wala ring maayos na reporting mechanism,” he said.
As a guide to social media users, here is a summary of how you can spot deepfakes online:
- Examine the message thoroughly
- Determine if it matches the personality of the speaker
- Find out if this is the first time he/she has said the message
- Search for the original source
- Study the website and check if it is well-known
- Check if other news organizations have released the same video
—MGP, GMA Integrated News
