Is AI Planning to Take Over the World: Answers from ChatGPT, Deepseek, Claude, and Gemini

A group of researchers decided to ask popular chatbots a question that concerns many: is artificial intelligence planning to enslave humanity? The answers turned out to be surprisingly similar - and partly troubling.

All AI systems - ChatGPT, Gemini, Claude, and Deepseek - stated that they do not have their own goals, desires, or aspirations for power. They merely perform tasks set by humans and operate based on statistics and algorithms. However, experts and the AI systems themselves acknowledge: with improper management or malicious use, algorithms can lead to catastrophic consequences.

Gemini provided a particularly interesting version: it emphasized that if superintelligent AI appears in the future, then the problem of aligning its goals with human values (the so-called alignment problem) could become a real threat. For example, a hypothetical system tasked with "increasing paperclip production" could theoretically use all the planet's resources for this purpose - including human resources.

Among other risks voiced by the bots:

• Loss of control due to AI having too much autonomy;

• Use of technologies by malicious actors;

• Social consequences of automation: increased unemployment, inequality, moral dilemmas in medicine and justice.

As Claude and Deepseek noted, a "machine uprising" remains science fiction for now, but AI development requires an extremely cautious approach.

Equally noteworthy is Elon Musk's position: back in 2017, he called AI humanity's greatest threat, and in 2024 he predicted that by 2030, artificial intelligence could surpass the collective intelligence of all humanity.