Wildberries, a leading digital platform in Eurasia, has launched an open beta version of its own deepfake detection service, designed to identify images generated using artificial intelligence.
The service enables users to determine, in just seconds, whether an image has been generated by AI. The beta version is free and available to everyone, with each user granted 100 image checks per month.
Following the open testing phase, the team will address any issues and determine the next steps for further development.
Future updates are expected to expand the tool’s capabilities, allowing for the detection of deepfakes in not only images but also audio, video, and text.
“In a world where the line between the real and the artificial is increasingly blurred, trust is more valuable than ever,” said Igor Somov, head of Trust & Safety at the united company Wildberries & Russ.
Somov added, “The beta version of our new service already enables anyone to verify the authenticity of any image, and in the future, this will extend to other media formats. We are proud to make cutting-edge technology accessible to all our users.”
The detector was developed in collaboration with professional AI artists, whose expertise, combined with the extensive experience of Wildberries’s internal team, has enabled the tool to achieve 95% accuracy in detecting generated images.
This service can prove invaluable in a range of scenarios, such as verifying the authenticity of images in news reports, confirming whether product photos in online reviews are genuine, and more.
Deepfake technology, which creates highly realistic synthetic videos and audios by altering faces, voices, and actions, is becoming more widespread.
While it creates new opportunities in fields like advertising and education, it also presents significant risks related to cybersecurity and the spread of misinformation.
Deepfake detection technology helps to mitigate these risks.
