The need for governments to adapt much more quickly to the changes brought by artificial intelligence, but also for a framework of global governance for AI to be built, was underlined by Kyriakos Pierrakakis, president of the Eurogroup and Minister of National Economy and Finance, in the context of the Athens Democracy Forum, which took place from September 29 to October 1 at the Athens Conservatoire.
Speaking with Thomas L. Friedman, columnist for The New York Times, in the discussion “Spotlight Conversation: A.I. vs. The World”, he stressed that “change is exponential, but our perception is linear”, as technological shifts that once took years or decades now unfold within months or even weeks.
The issue, as he said, is to harness AI’s enormous potential for productivity, growth and innovation, without underestimating the risks.
“We focus more on vocabulary than on implementation”, he noted, arguing that today’s institutions are not yet properly designed for the age of artificial intelligence. “Institutions are the software of our societies. And this software must change in order to adapt to the AI economy”.
He also referred to the experience of the digital transformation of the Greek state, stressing that many governments have not yet completed even the previous stage, that of digitization.
Interoperability, the use of data and digital transformation are, as he said, necessary prerequisites for states to be able to harness AI in health, education and services to citizens.
He also argued that policymaking can no longer be done only by economists and lawyers. “We also need people in the room who understand technology”.
He placed particular emphasis on cybersecurity. AI, he noted, “can identify system vulnerabilities with unprecedented speed and scale, while states around the world continue to operate with old digital infrastructures”.
One of the responses being proposed “is to use AI to protect against AI”, but this must be accompanied by strong governance rules.
“No state in the world is large enough to manage artificial intelligence on its own”, he stressed. Europe, he added, must be at the table where decisions are made and cannot be confined to the role of regulator.
“AI is not a field from which Europe can afford to be absent”. He even predicted that in the coming years European companies will emerge that will grow very quickly in this particular field.
He pointed out that the agenda of economic policy itself has already changed. “If you are a finance minister today and you are not discussing energy, AI and defense, you are not doing your job properly”, he said, describing artificial intelligence as one of the major challenges of today’s generation of politicians.
The same, as he stressed, also applies to education, raising the question that now concerns education ministers internationally: “How do you design an education system for a world that does not yet exist?”.
The answer “does not lie in learning a specific programming language, but in the development of algorithmic thinking, logic, imagination, collaboration, communication and the ability to listen to others”.
In the same discussion, Thomas L. Friedman described AI as a “phase change” corresponding to those in physics: from the “ice” of the age of tools, to the “water” of the age of information and now to the “steam” of the age of intelligence.
“We are one-tenth of a degree away from the greatest phase change in human history and we have no idea what is around the corner”, he said. He warned in particular about powerful Chinese open-source models, which can be modified, have their safety guardrails removed, and operate even with limited computing infrastructure.
He argued that AI should no longer be treated simply as just another technology. “We have created a new species”, he said, using the image of a new “migration” of AI agents into human societies: the question, as he put it, is how we will learn to live with them without “becoming their pets”.
He also insisted that the US and China should create common structures for governance, authentication and building trust around AI.
“Is it naive to believe that the US and China can cooperate? Yes. There is only one thing more naive: believing that either of the two will be fine if they do not cooperate”.
The discussion was moderated by Patrick Healy, assistant managing editor for Trust and Standards at The New York Times.