How AI fakes are making the world more confusing than ever

Thanh Thu first encountered an AI hoax last month when she believed a photo showing dozens of people clinging to rooftops during flooding in Thai Nguyen Province in the north was real. The grainy quality made the picture convincing at first glance.

"But on closer look, it made no sense as multiple helicopters were hovering just above the water yet not rescuing anyone," she says

Thanh Thu, 28, from Ho Chi Minh City, looks over alerts about scammers using AI-generated photos and videos to steal money, October 2025. Photo by VnExpress/Nga Thanh

That experience made her more cautious. So when she saw the new flooding video, she examined it again and noticed the grandmother’s face was blurred and distorted, while nearby people appeared calm, chatting casually with water only reaching their knees.

"The footage was heart-breaking at first and gave me a strong urge to share it, but I realized in time that it was AI-generated just to attract views," the 28-year-old in HCMC says.

In recent months, her social media feeds have been filled with photos of friends posing in destinations around the world, their faces retouched to look like models. But when she noticed that photos from various friends had identical settings, with only their faces changed, she realized the images were AI-generated.

In Hanoi, Hoang Viet, 27, has also been fooled by AI-generated content. A nature enthusiast, he once spent more than an hour watching a TikTok video about a supposed "rainbow mushroom" and even took notes for research. "Later I did more research and asked experts only to find out the species does not exist at all," he says.

He recalls being frightened by clips of "sea monsters revived" or wild animals jumping on trampolines, all low-quality videos resembling security footage but drawing tens of millions of views. "All of them were fake." After these experiences, he now cross-checks any information using reliable news sources or academic materials before accepting it as true.

Screenshot of a video generated by AI. Photo by VnExpress/Nga Thanh

The rise of tools such as OpenAI’s Sora, Google’s Veo 3 and Runway this year has caused a surge in AI-generated images and videos, from staged travel photos to recreated images with deceased relatives.

According to digital transformation expert Dr. Dinh Ngoc Son, AI-generated content is designed to exploit emotions and curiosity. "Sensational or unusual information always spreads faster," he says. "As AI improves realism, the spread becomes stronger, creating an emotional whirlwind that is difficult to control."

He says authenticity and digital ethics have become major challenges. As generative AI advances, the line between real and fake is becoming increasingly thin. He cites fabricated flood rescue clips and deepfake news anchors as examples.

In July Malaysian newspaper The Star reported that a couple in Kuala Lumpur drove 300 kilometers to a tourist destination only to find it did not exist and had been entirely produced by AI. "When users share such content without verification, it distorts public perception and undermines trustworthy information platforms," Son says.

Cybersecurity specialist Ngo Minh Hieu, known as Hieu PC, says many people unknowingly expose personal data by uploading their photos to AI-powered editing apps. "With just a smartphone, anyone can create AI products," the founder of the "Anti-Scam" project says. "While many use them for fun, scammers increasingly exploit them to deceive those who are less tech-savvy."

At a national seminar on online fraud prevention in late October, Deputy Minister of Public Security Pham The Tung warned that criminals are taking advantage of AI and deepfake technologies to defraud citizens. On Oct. 28 anti-scam platform Chongluadao.vn issued a new alert about fraudsters using AI to manipulate e-commerce livestreams.

They take real sales videos, modify the sellers’ appearances to look physically disabled or ill, and re-upload them to solicit purchases from sympathizers. Scammers also create dozens of fake accounts to leave emotional comments, leveraging the bandwagon effect to deceive more viewers.

As someone who once bought products on livestreams of disabled people out of sympathy, Thu says the AI scam leaves her increasingly bewildered. "Now, when even images of those people can be faked, I do not know what to believe."

To combat AI-driven misinformation, Son says joint action is needed from individuals, society and authorities. Individuals must improve their fact-checking skills, while the journalism and education industries should strengthen digital media literacy, he says.

"The foundation remains digital ethics. When each person is equipped with a personal ‘filter,’ AI will no longer be a threat."

Hieu adds that Vietnam’s upcoming Artificial Intelligence Law, which includes personal data protection provisions and is expected early next year, will help safeguard both users and businesses.

Thu now carefully reviews every piece of content before sharing it. "When truth and falsehood are nearly indistinguishable, I must spend extra time verifying, so I will not be led by those chasing clicks or scams."