Paris – February 11, 2025: French President Emmanuel Macron (before at the center) poses for a group photo with world leaders and participants (including Dr. Alondra Nelson) at the end of the plenary session of the summit of action AI at the Grand Palais. (Photo of Ludovic Marin / AFP via Getty Images)
At the top of the AI action of Paris from February 10 to 11, 2025, France and India co -sponsored A rally of heads of state and government, leaders of international organizations, CEO of small and large companies, academics, representatives of non -governmental organizations, artists and civil society members for the third in a series of international discussions On political questions surrounding artificial intelligence (AI). The main themes of the summit were the public interest of the AI, the future of work, innovation and culture, openness and confidence in AI and global AI governance.
The summit ended with the American vice-president JD Vance first speech on the international scene. Vance promised the American domination of AI and warned other countries, in particular the European Union, against the “tightening of the screws”.
“The future of AI will not be won by hand security,” said Vance. “It will be won by the construction, from reliable power plants to manufacturing facilities that can produce the fleas of the future.”
In particular, the United States and the United Kingdom have not signed the summit Final press releaseagreed by 60 nations, including France, India and China. It was the first time that the two countries were not signed by such global statements in the Summit series, which launched Bletchley Park in the United Kingdom in 2023 and gathered in Scramble in 2024.
The vice-president Vance was one of the world leaders, CEOs and a selected audience present at night before delivering his speech during a private “work dinner” organized by French president Emmanuel Macron at the Palais de l’Elysée. US teacher Alondra Nelson was one of the guest speakers.
Dr. Nelson holds the Harold F. Linder chair at the Institute of Advanced Study. She was a member of the Summit AI AI Action Committee and participating in the IA 2023 security summit organized by the United Kingdom in Bletchley Park. Dr. Nelson served the United Nations High -level advisory body on artificial intelligence. She is a former interim director of the Bureau and Technological Policy Office of the White House, where she led the development of the Bide-Harris administration “”BluePrint for a declaration of AI rights“, Which served as the cornerstone of the executive decree of President Biden on safe, secure and trustworthy development and the use of artificial intelligence, which was canceled by the Trump administration.
Below are its comments in their entirety. -Jh
Three errors
Remarks to Elysée Palais on the occasion of the AI action of AI action
Alondra Nelson
Paris, France
February 10, 2025
Mr. President, Mr. Vice-President, Excellency, Distinguished Guests, Women and Gentlemen:
Allow me to express my gratitude to President Macron for welcoming this vital summit and exercise my appreciation to Anne Bouverot, whose leadership as a special;
This evening I would like to talk about three mistakes – three fundamental false ideas in the way we think of artificial intelligence. These are ideas on AI, many of us have heard and internalized – in our media, our politics and in our business – which no longer serve us well.
These are errors in good faith, stories and conventions that have emerged while our societies have tried to give meaning to this rapid technology. To what extent we are lucky to have this space to engage these false ideas and together established a better course for the future.
The first error is that the purpose of AI is efficiency and scale.
We have heard that the purpose of AI is to accelerate growth, to compete in the world market, to improve the pace of productivity.
And indeed: these are AI functions.
But we must never confuse the function for purposes. We must not confuse what AI can do with whom the AI should serve.
The purpose of this power should be to benefit humanity.
To make life better and easier for people. To improve health and facilitate suffering, protect security and promote security. To support the dignity and opening of opportunities, and amplify the complete spectrum of human potential.
The purpose of AI is not the scale or efficiency. The goal of AI is people.
The second error we heard is that AI requires a compromise – between security and progress, between competition and collaboration, and between rights and innovation.
But each of them is a false choice.
History shows us that real innovation is stimulated by thoughtful governance. Environmental standards have led the development of catalytic converters. The aviation regulations have given us safer planes and brought our savings closer.
With a scientific collaboration and a fluid market of ideas, we have seen rapid progress in the folding of proteins and the modeling of the climate, and much more. When researchers can openly share methods and models, when developers can rely on everyone’s work, we create a virtuous innovation cycle.
Supporting research to align AI with our values gave us clearer methods and better models. To reach the breakthroughs forces us to reject the false binary. He calls us to govern thoughtfully.
The third error, and the last one I will share, is as follows:
It is not inevitable for AI to lead to great public advantages. The results that many of us hope or anticipate are not characteristics inherent in technology itself. They are not coded in algorithms or integrated into neural networks.
The good that can come from AI, science and education, in health care and in the economy, must be actively managed by leadership and real innovation – innovation which is fundamentally responsible and centered on man.
These advantages will not only emerge from the invisible hand of market dynamics. They must be cultivated in partnership with civil society and with our democratic institutions. And above all, AI in the public interest requires the significant involvement of the same people whose life could be transformed by these technologies.
I am sober about threats to artificial intelligence, how this technology can perpetuate discrimination, threaten security and disrupt social cohesion through continents.
But I’m finishing this evening with a word of hope:
The print press has not only printed books – it has democratized knowledge. The phone has not only transmitted the voice – it connected families over large distances. The Internet has more than linked computers – he created unprecedented opportunities for collaboration and creativity.
We have the tools to guide AI to work for all our employees.
If we go beyond these three errors, if we advance thoughtful governance, we can guarantee that the AI systems improve rather than reducing human rights and dignity.
We can create systems that develop the opportunity rather than concentrate power. We can build a technology that strengthens democracy rather than undermining it.
Now, while we feel what can be the most transforming innovation in humanity, we must make sure that AI follows this tradition.
THANK YOU.
