AI Data has built an in-house verification environment for an AI video identification platform using ”SYNTHETIQ VISION,” a deep fake detection AI that is being researched and developed by the National Institute of Informatics (NII), and has begun evaluating its effectiveness and operational methods.
At the moment, it is only in the verification phase towards actual operation and service provision.
We mainly examine judgment trends for various sample videos assuming crimes and trouble cases. Comparisons are also being made with conventional visual and manual confirmation processes. In addition, the company is also verifying its legal adequacy and potential for incorporation into internal controls and audit processes.
If AI video appraisal infrastructure is put into practical use, it will be possible to screen for possible deepfakes, such as defamatory videos in which faces of children and students are synthesized without permission, and pornographic content. It will also be possible to mechanically extract videos suspected of being ”spoofed” or ”synthesized faces.” You can also visualize the risk of damage to your own brand and the spread of false information.

※Translating Japanese articles into English with AI