Those who adapt faster will survive: AI enthusiast Krakovetsky answers whether professions will disappear
29 April 16:59
The era of AI-first is upon us: autonomous agents will make decisions without humans, scientific articles will be written by AI, and businesses and states will compete for control over clean data rather than investments. Why AI content checkers no longer work, whether Google penalizes texts created by AI, and how not to lose the technological race – read in the article
At the Discussion Panel “AI in Business Communications and PR. What are the benefits and where are the hidden threats?” organized by
We are already on the verge of the “AI-first” era
Krakovetsky described the current state of technology development as the era of AI-assistant, when AI helps humans to perform tasks faster and more efficiently. However, according to him, this is only the beginning.
“In the next 2-5 years, we will enter the AI-first era, when artificial intelligence will become the main decision-maker. Humans will remain, but their role will change: they will check, guide or approve final decisions,” Oleksandr Krakovetskyi
According to the speaker, major players such as OpenAI and Microsoft are already actively working on the creation of “agents” – autonomous AI systems capable of performing complex tasks independently.
AI is already writing scientific papers better than humans
One of the most striking examples of rapid progress is a startup from Japan that has created an AI system capable of conducting scientific research on its own.
“Artificial intelligence chooses a topic, researches it, writes code, analyzes the results, draws conclusions, and writes an article that successfully passes peer review and is rated better than human work,” Krakovetsky emphasized.
This fact, in his opinion, shows that we are already on the verge of a radical transformation of scientific and creative activity.
Another important point that Krakovetsky drew attention to is the tremendous growth in the intellectual capabilities of models:
“The benchmarks that were supposed to evaluate the development of AI have simply ended. Models have passed them. Now we are forced to create new benchmarks because the old ones do not work anymore,” Oleksandr Krakovetskyi
For example, he said that modern language models have already crossed the 130 IQ score line, and this figure is growing rapidly.
The progress in solving complex math problems is particularly impressive:
“Six months ago, the accuracy of Frontier Math was 3%. Now it is already 24%. And on another test, the result grew from 13% to 83% in just 4 months,” Oleksandr Krakovetskyi
Does Google penalize content created by AI?
Krakovetsky also dispelled a widespread myth about the “pessimization” of AI content at Google.
“Google officially states that it does not pessimize content created by artificial intelligence. They penalize inaccurate or spammy content, but high-quality text,” Oleksandr Krakovetskyi said
At the same time, he emphasized that existing tools for detecting AI content are unreliable. According to him, checkers that determine whether a text is written by AI produce a large number of false positives. They often mistakenly identify authentic human texts as created by AI just because they are structured.
Professions will not disappear but will change
The speaker paid special attention to the future of professions:
“Professions will not disappear. They will transform. And the most important skill of the future is the ability to adapt quickly, to move from one format to another,” says Oleksandr Krakovetskyi
In his opinion, it is worth less “shaking” over the author’s texts and more work on developing expertise and flexibility. Mr. Krakovetsky also drew attention to a really critical problem:
“Muslims are engaged in data poisoning – they create fake data that poisons large language models. We need to open archives and create our own clean datasets, because the future belongs to those who control AI training,” Oleksandr Krakovetskyi
According to him, any modern organization should become an IT company, building an infrastructure based on its own data and AI. Summarizing, Krakovetsky added that we all have a chance now: either to lose this technological race or to see global trends and move forward. The main thing is not to get hung up on trifles but to work ahead.
