Kazakhstan is developing an AI-based system to detect deepfakes

“We are creating a digital initiative for the automatic detection of fake content and the identification of deepfakes. This technology is based on artificial intelligence and has already demonstrated its effectiveness in combating online drug trafficking. For example, the AI-powered ‘MERGEN’ system automatically monitors the information space, identifies online drug stores through direct and indirect signs, and blocks them. This week the 5,000th such resource was shut down. We are now training an LLM model to include functions for detecting deepfakes,” Kochetov said.

“The program was developed by our own team, domestic programmers. This technology is already being successfully applied. But this program is only an assistant. It identifies potential trigger points in large volumes of data, and then representatives of legal monitoring review them. Therefore, there will be no accidental blocking — that is certain.”

“We are working on it. I think, not two years. We will, let’s say, understand something fully within the next few months. This is not a quick process, because it is training. A very large amount of content needs to be fed in. It is daily work. The information and digital unit of our ministry is aimed at this, to at least keep pace with the times and, in some areas, move ahead. So we will try, we will catch up. If our opponents are in daily search of new schemes, then we are in daily search of new solutions that will allow us to neutralize those schemes,” Kochetov emphasized.