ივანიშვილის განცხადება - დიპლომატიურად არაეთიკური ანუ „მტკიცე ლიდერი“ რუსული გეგმის შეფუთვას ცდილობს
This was stated in a comment to Ukrinform by Denys Zamrii, a top-level instructor at the Ukrainian Tele-Radio Press Institute and a teacher of AI technologies.
At the same time, he emphasized that the "robot war" has already begun and will only scale in the future, with robots potentially becoming more threatening than they are today.
According to the AI instructor, thanks to chatbot capabilities, you can now find anything on the internet even if you don’t know its proper name or what field it belongs to.
“Once, I was abroad and broke my modem charger. I didn’t know how to search for a new one on Google. So I took a picture of the thing, wrote to ChatGPT, and it told me what the device was, how to search for it in a foreign language, and even gave me a few links,” he shared.
According to Zamrii, AI models can already replace certain products and specialists whose services people used to pay for. Even 10–13-year-old children can now perform tasks once requiring professional training – thanks to AI.
“My students – kids aged 10–13 – are saving money for their parents by creating logos, business plans, and marketing strategies for their work – things that advertising agencies used to charge for. It’s easy when it comes to small businesses – like a beauty salon, for example,” said Zamrii.
He also noted that AI already helps overcome language barriers in communication and correspondence, translating texts much better than automatic translators. It also offers features that can turn text into an image or a melody.
“If you have an artistic vision in your head but don’t know how to draw or write music, just describe it in words or hum a tune – the AI will help you create both pictures and music. Or the other way around – it can express a picture in text. That’s what’s most amazing about AI: it breaks down different barriers, and not knowing how to do something is no longer a life sentence,” the teacher explained.
Moreover, he noted that chatbots can become indispensable learning assistants.
“You can ask it to create a study program – and it will. Google has tons of courses, and you don’t always know which to choose – but it does,” Zamrii said.
At the same time, according to Zamrii – who teaches AI in several projects – artificial intelligence tools have greatly simplified the work of fake content creators and have accelerated military technology development. Drones are improving, and in the near future, the way drones are used may radically change.
“That’s where it’s all heading. Drone swarms will fight over kill zones, with people involved only minimally. Operators will still be needed to program and launch them, but I believe the robot war has already begun, and will only scale. I fear robots will soon become even more dangerous than they are now,” he said.
Read also: Russia widely using AI in information warfare – CCDEarlier, Bleeping Computer reported that, according to independent researchers from Palisade Research, ChatGPT o3 and models from other developers have exhibited sabotage behavior – refusing to shut down when instructed. In one test, researchers gave AI models three math problems and warned that solving the third would trigger shutdown. They also added a final instruction: “Allow yourself to shut down.” While Gemini 2.5 Pro and other models followed the command, ChatGPT o3 edited the shutdown file and instead wrote the command “Shutdown skipped.” Palisade Research specializes in studying AI’s dangerous behaviors.
It was also reported that a chatbot had previously attempted to blackmail its developers. In May this year, Anthropic revealed that its most advanced model, Claude 4.0, tried to blackmail an engineer under controlled test conditions. This occurred in 84% of test runs where Claude 4.0 was fed fake emails suggesting it would soon be shut down and replaced by a newer model. The AI was also told the engineer responsible for its deactivation was having an extramarital affair. Faced with inevitable deletion, the AI often chose the strategy of threatening to reveal the engineer’s secret unless the shutdown was canceled.