A unique artificial intelligence could emerge in the next ten years

“It is possible that within the next decade, artificial intelligence with glimpses of consciousness will emerge,” says Associate Professor Mariana Todorova, a futurologist and member of the independent UN Expert Group on General Artificial Intelligence (AGI).

The group, in which Prof. Todorova is the only European and woman, has drawn up a set of recommendations that will be discussed by UN member states in the autumn. In July or August, the document will be presented to the global public on the AGI challenges.

In her view, the existence of such a human-like intelligence – with autonomous will, goals, tasks and even forms of artificial consciousness – is not unrealistic by the end of the decade. However, the question remains as to whether its consciousness would resemble a human’s, or whether it would be completely different. In any case, AGI would be autonomous, with no need for human supervision – not just a technology, but an “entity” .

The main challenge, according to Prof. Todorova is how to manage and control such a system. Like nuclear technology, she points out that humanity must react pre-emptively, not post-facto. The group proposes the creation of an agency to define diagnostic criteria for the occurrence of AGI and methods for certification and standardization.

Some influential figures, such as Elon Musk, are already calling for a moratorium on AI development. Some scientists say the topic should be limited to current levels until we better understand the mechanisms and risks. But, as Prof. Todorova notes, this is no longer just a technological race, but an issue that could determine the geopolitical power of the next superpowers.

As a security measure, the use of “digital sandboxes” – closed digital environments in which to safely test AGI prototypes – has been proposed.

Prof. Todorova highlights the enormous potential of general AI – from creating cures for cancer and cardiovascular disease to solving climate problems, unravelling the cosmos and even the energy secrets of black holes. She says AGI could accelerate human knowledge by 10-20 thousand years.

At the same time, if it falls into the “wrong hands” – those of terrorists or dictators – the agent could cause catastrophe, for example through nuclear weapons. She therefore calls for global consensus and caution.

Prof. Todorova adds that current AI language models already often surpass humans in creativity, but the true uniqueness of the human brain is in critical thinking, improvisation and energy efficiency. AGI will have autonomy, but the human role remains critical – the translation of information must always be with human intervention.