
The benefits and also dangers of “agentic artificial intelligence” at a time when Europe is striving to become more competitive in artificial intelligence were the focus of a discussion between Digital Governance Minister Dimitris Papastergiou, acclaimed scientist Francis Fukuyama, and experts in the field on Wednesday, at the Economist’s 29th Annual Government Roundtable in Athens.
Papastergiou stressed that what we have known of AI up until now is its use of data and “what is has learned to do, and it does [things] more quickly and possibly more accurately than human beings. Therefore, the first thing we must see is what data we are supplying to AI.”
Fukuyama referred to the dangers presented by so-called agentic AI that is capable of making autonomous decisions, noting that there were two kinds.
One was the use of this technology by “bad people for evil ends,” noting that this was a process that was ongoing and included such things as deepfakes, hacking, criminal scams and election rigging.
The second, and according to Fukuyama more dangerous, was not linked to generative AI but agentic AI, which takes initiatives, and the risk of making the right decisions.