The fear of artificial intelligence is something that comes from afar. From long before OpenAI developed ChatGPT, in fact. You just need to take a look at movies like Terminator o Matrix (among others, they are simply the most famous) to realize this. However, for some time now more and more experts are talking about the threat.
Among them, Daniel Kokotajlo. Although this researcher is not as high profile as Sam Altman, he was actually his colleague, and one of the top managers of his company. Since he left it, however, he has only warned of the enormous danger that artificial intelligence poses. So much so that It has even set a date for the destruction of humanity.
AI, the worst enemy of humanity
Until recently, it seemed like apocalyptic fears around AI had calmed down. The catastrophic predictions, in fact, had been largely replaced by problems much closer to the people, such as massive layoffs in many companies, or security breaches. Problems even recognized by Sam Altman and OpenAI.
But on the other hand, Daniel Kokotajlo continues in his thirteenth. What’s more, this expert in artificial intelligence has been convinced that AI will evolve, and that when it does humans will be at enormous risk. The good news, if it can be considered such a thing, is that now the specialist has given a little more time before the machine revolution occurs.
As reported by The Guardian, Kokotajlo presented a speculative model in April 2025 in which he proposed that AI could completely automate programming by 2027, which, according to his scenario, would lead to an explosion of artificial intelligence and possibly an existential risk for humanity. That would be right around the corner.
In new statements, however, he has postponed his final sentence until a little later. Now he and his team estimate that those milestones they probably won’t occur until the early 2030s, with a possible arrival of «superintelligence» around 2034. Apparently the development of AI is slower than they had anticipated.
The 1930s, a prodigious decade
In this way, the former OpenAI employee’s perspective matches those of many other experts, including Elon Musk himself. Many are those who have already marked 2030 (year up, year down) on the calendar, and will then have the champagne ready to welcome artificial superintelligence or AGI.
Does everyone believe that, for better or worse? Not really. Some computer scientists consider that the evolution of AI is nothing short of nonsense. But is it true that when the river sounds, it carries water? Kokotajlo, of course, is convinced of the danger.