A New AI-Driven Photo App Wows China, But Privacy Fears Loom
Offering high-quality professional photos at remarkably low prices, a new AI-powered portrait generator mini program has emerged as a potential disruptor in China’s photography market.
Miaoya Camera, a mini program available on WeChat, allows users to upload a single front view headshot and a set of 20 or more recent photographs. In return, they receive a collection of AI-generated half-body portraits showcasing various angles, lighting, and facial expressions.
While the program has taken China’s youth by storm since its release on July 17, with thousands rushing to put it through its paces, some netizens have raised concerns about potential privacy issues.
Priced at 9.90 yuan ($1.38), the service has swiftly gained popularity. During peak hours, some users reported enduring wait times of more than 10 hours. When Sixth Tone tested the program Wednesday, nearly 3,000 users were already in line, which meant a wait time of around five hours.
Since its recent release, thousands of users have shared portraits generated by Miaoya Camera on popular social apps, lauding the quality it offered. In a viral post on the lifestyle platform Xiaohongshu, a poll indicated that 72% of users preferred the profile photo created by the mini program over one taken at Tianzhenlan, a major professional photography chain.
“You can recognize it’s you since the photo is not overly filtered, but it’s a much better version,” a woman surnamed Liu from Beijing told Sixth Tone. She added that she and others have already opted to use the AI-generated portraits from Miaoya Camera as their profile photos on LinkedIn.
According to Miaoya Camera’s previous user agreement, users would be “enabling us (the company) to use your content in any form, in any media or technology (now and in the future).”
In response to user concerns, Miaoya Camera made revisions to its user agreement on July 20.
The updated agreement now assures users that the photos they upload will solely be used for creating digital avatars. Additionally, the company said it would not extract or employ the photos for identification or any other purposes.
Song Yuhao, an independent security researcher, told The Paper, Sixth Tone’s sister publication, that a data leak resulting from an external attack or internal misconduct could pose serious concerns. “Users’ photos and associated information could potentially end up on the black market, leading to the risk of targeted fraud or other criminal activities,” he said.
“If users’ photos are used for model training without their knowledge, some of their facial features may appear in images generated elsewhere, which they may not want.”
The increasing use of AI-generated content, particularly deepfake technologies, has raised significant concerns about personal privacy and data protection in China.
To address this growing threat, China has taken steps to strengthen its regulatory framework in recent months. A major guideline on generative AI technology, scheduled to take effect on Aug. 15, mandates service providers to obtain user consent for data processing activities involving personal information.
Additionally, service providers are required to respond to users’ requests for data deletion and other uses of their personal data.
Additional reporting: Jiang Zuer; editor: Apurva.
(Header image: Promotional photos of Miaoya Camera. From Weibo)