AI photo users believe personal data not worth protecting

Anh, 25, said even before all the AI craze broke out in Vietnam, she and many friends on social media had uploaded their own photos to applications to modify them, before re-uploading those photos online. As such, she believes her own photos have long been harvested by different applications, and does not care how the parties would utilize such data.

"I used to be worried seeing warnings about facial deep fakes. But I have already shared my information to dozens of applications for many years, so maybe my facial data can no longer be protected on the Internet," she said.

In recent times, AI-generative applications have been sprouting up as an emerging trend, where many people use them to generate their own photos. Over the past few days, social media users in Vietnam have been sharing photos created by AI Avatar, which is a part of the Zalo messaging app, apparently with little concern for personal data security.

AI Avatar requests access to users' cameras or photo albums, and asks for facial photos with clear features. Users would also need to enter certain information, like their gender, age and the photo style they want.

Besides AI Avatar, popular similar AI applications in Vietnam include Lensa, Loopsie and Reface.

The representative of an anti-scamming project in Vietnam said not everyone carefully reads the terms and conditions of all services.

"Users may use the app, but that also means agreeing to its terms, including the fact that said services can include terms that cause disadvantages to users in the future," they said.

Vu Ngoc Son, technology director of the Vietnam National Cyber Security Technology Corporation, said while users only carry out simple steps like uploading and downloading photos, the actions still harbor risks, as the uploaded photos are still stored within the servers of service providers.

"Concentrating all the photos in one place harbors risks of getting attacked by hackers. If the photos fall into the wrong hands, they can use deepfakes to create fake photos and videos for different purposes, including scamming," Son said.

As deepfakes are getting more frequent online, the sharing of personal photos often warrants warnings from experts. But many are negligent about the matter.

"Dozens of apps on the phone have already requested access to users' photo libraries. Would protecting oneself from a single app be meaningful?" Nguyen Hung asked on one tech forum.

The anti-scamming project representative said the inquiry on the forum is "partly right", but it shows the lack of responsibility over one’s personal data.

"Such a way of thinking is negligence, making one less able to protect themselves online. Once your data is exploited, the ones who have to lose are not just one user, but also their families, friends and everyone around them," they said.

Son said: "You might have had your data leaked before, but that’s old data and may no longer be accurate in the present. Moreover, data collected through third parties would be less reliable and accurate. As such, entities that want to exploit data will always want the latest data collected in the most direct way possible."

Zalo has yet to provide a comment on how the platform protects its users' data.

Experts recommend people to only provide information that is truly necessary, as well as carefully checking the terms and conditions of services. For photo modification applications, users can utilize protective measures integrated with the phones’ operative systems, such as not allowing full access to photo libraries.