OpenAI chief shocks: wants superintelligent AI in a few years

Sam Altman, CEO of OpenAI, made a surprising statement at the beginning of 2025. According to him, the company already knows how to create general artificial intelligence (AGI) and is now aiming even further – towards superintelligence.

Superintelligence within reach?

Altman estimates that superintelligent AI systems could become a reality in just a few years. These systems, according to him, should accelerate scientific research and innovation.

An interesting perspective is the definition of AGI according to the agreement between OpenAI and Microsoft. According to it, these are systems capable of generating at least 100 billion dollars in profit. When OpenAI reaches this milestone, Microsoft will lose access to its technology.

AI safety remains unresolved

A disconcerting fact is that OpenAI itself admits it has no solution for controlling superintelligent AI. At a time when the company is aiming to create such advanced systems, it has paradoxically dissolved teams focused on AI safety.

Despite all the ambitious plans, today's AI systems still have significant flaws. They hallucinate, make fundamental errors, and their operation is costly. However, Altman believes that these obstacles can be quickly overcome. The history of AI development, however, shows that time estimates are often overly optimistic.

According to Altman, AI agents – autonomous systems capable of independently completing certain tasks – could enter the workforce as early as 2025. This could fundamentally change corporate operations and work productivity. The question remains, however, how society will prepare for this change.