«Sber presented the GigaChek technology, which helps determine the origin of the text: whether it was written by a person or generated by artificial intelligence. You can try the solution in the demo version of the service on the website or through the chat bot in Telegram.
Fonte da imagem: sber.ru
The service that helps determine the origin of the text will be useful when checking diplomas, dissertations and other scientific works; it will be used by editors who accept the work of copywriters and writers; and owners of public pages will be able to identify comments written by neural networks. For now, the text is assessed based on the content that prevails in it. In the future, the technology will be improved and will allow interval search – such an opportunity should appear in the foreseeable future.
«The goal of our technology is to have such a level of quality of work that in order to bypass it, the text will have to be distorted to the point of inapplicability to real tasks. For example, a diploma with scattered symbols “;” will simply not pass the quality control, and such a comment in the public will immediately be highlighted,” Sber said. When the new service is officially launched, the company will prepare API tools that will help integrate the technology.