Replay the session and read a summary below.

HOW TO GOVERN ARTIFICIAL INTELLIGENCE

Panelists:

• Marc Mossé, Senior Counsel, August Debouzy Law Firm
• Gabriela Ramos, Assistant Director-General, UNESCO
• William Yee, International Governance and Diplomacy, PSIA

Moderated by:

• Beatriz Botero Arcila, Assistant Professor of Law, Sciences Po

Introduced by:

• Ricarda Seitz, International Governance and Diplomacy, PSIA

“Raise your hands if you believe it’s possible to have AI regulation without human intervention” – it was with these words that Ricarda Seitz opened the AI panel taking place in Chapsal’s amphitheater. Across the room, only a handful of people lifted their hands, others looking more skeptical.

At a time when AI is making headlines, thanks to the success of new OpenAI tool “Chat GPT”, the discussion that was about to take place seemed more than necessary to all those attending. Indeed, as Ricarda Seitz then said, these new technologies and algorithms “have severe implications for our privacy, security, and ultimately the way we learn, work, and fight for our democracies.”

Where do we go from here?

AI is already present in our reality, it is no longer a dystopian prediction of our distant future, and since we can’t put the genie back in the bottle, we are left with regulation.

“We can’t control everything that will be done. But we can establish rules”, said Gabriela Ramos, who presided over the “Ethics of Artificial Intelligence” – the very first global agreement on the ethics of AI to be adopted by the UNESCO’s 193 member states.

As such, Gabriela Ramos addressed her concerns on allocating responsibilities in the digital world, not taking things for granted, and how urgent it is to reflect all these challenges into “effective policies”. To these statements, William Yee added that as of right now, policies are mostly a national concern. However, he said, “AI will have transnational effects”, so universal schemes must be taken more seriously in the regulatory process. Existing regulations, Ramos also claimed, are insufficient. “GDPR is looking at privacy, but it’s not looking at fairness”

Fighting discrimination 

“AI is just amplifying what we have done in the past with our pre-existing biases” – this powerful quote was brought out by William Yee.

This statement was corroborated by Gabriela Ramos, who proceeded to mention several examples to illustrate the issue: “We don’t realize that what is behind that the business model of these companies is non inclusive, there is much cultural diversity being lost. If you’re not represented in the data, AI has a hard time recognizing and including people in its algorithm.” Among the misrepresented, are women. “It is a white man’s world,” Gabriela Ramos added. Overrepresentations are also a risk. “If you run an algorithm on crime prediction, it [AI] will always tell you that the one who will commit a crime will be a black man, because they are over-represented in the data.”

As a solution, Marc Mossé said that AI must be infused with a set of principles: fairness, safety & reliability, privacy & security, inclusiveness, and accountability.

Combining technology and social sciences

“More than a multilateral approach, the civil society is part of this multi-stakeholder approach,” said Marc Mossé.

In a world mostly led by tech experts and developers, the panel agreed on the importance of a multilateral and multidisciplinary approach to AI. Only that way will it be possible to be optimistic about technology.

When answering the question about who should be in charge, Gabriela Ramos raised an essencial point: “This is not a technological debate; it is a societal transformation.”

(c) An article by Mariana Abreu, a student in the Joint Master in Journalism and International Affairs.

More information about the Youth & Leaders Summit 2023.