Humanity is in a race against time to learn how to harness artificial intelligence (AI) for the greater good while avoiding the dire risks it poses, a senior UN official and experts said on Thursday .
“We have let the genie out of the bottle,” said Doreen Bogdan-Martin, director of the United Nations’ International Telecommunications Union (ITU).
“We are in a race against time,” she said at the opening of a two-day global summit titled AI for Good [l’intelligence artificielle au service du bien] in Geneva.
“Recent developments in AI are simply extraordinary,” she acknowledged.
Thousands of conference attendees learned that advances in generative AI are already accelerating efforts to solve some of the planet’s most pressing problems, such as climate change, hunger and welfare.
“I think we have a once-in-a-generation opportunity to put AI at the service of everyone on the planet,” Bogdan-Martin said in an email sent to AFP before the top.
But she deplored Thursday that a third of humanity still remains completely disconnected and is “excluded from the AI revolution without being able to express itself”.
Concentrated power
Ms. Bogdan-Martin stressed that AI holds “immense potential, both for good and evil,” and that it is essential to “secure AI systems.”
She said this was particularly important today, given that “2024 is the biggest election year in history,” with votes in dozens of countries, including the United States.
However, “with the rise of sophisticated disinformation campaigns of the ‘deep fake’ type, it is also the most controversial year,” she added.
“Not only does this misuse of AI threaten democracy, but it also endangers the mental health of young people and compromises cybersecurity,” warns this official.
In a speech at another AI governance event this week, the ITU chief warned that “the power of AI is concentrated in the hands of of too few.”
“It’s risky and ethically fragile to be in this kind of position for humanity,” she said.
Other experts present at the conference agreed with this.
“We need to understand where we are heading,” said Tristan Harris, a technology ethicist who co-founded the Center for Humane Technology.
He recalled the lessons learned from social media, which was initially presented as a way to connect people and give everyone a voice, but which also spawned addiction, viral misinformation, online harassment and an explosion in mental health problems among adolescents. .
” Super powers “
While AI can benefit humanity in many ways, Harris warned that the incentives driving companies to deploy the technology risk greatly increasing these negative effects.
“The main motivation for Open AI or Google’s behavior is the race for market dominance,” he said.
In such a world, he added, “governance that evolves at the speed of technology” is vital.
OpenAI director Sam Altman, who rose to prominence worldwide after the release of ChatGPT in 2022, acknowledged that this technological advancement could carry dangers.
Speaking via video link, he told the gathering that “cybersecurity” was currently the biggest concern regarding the potential negative effects of AI.
However, in the longer term, he said it would probably be necessary to “modify the social contract, given the expected power of this technology”.
Overall, however, he insisted that, from the perspective of the historical evolution of new technologies, AI systems were “generally considered safe and robust.”
While welcoming discussions on regulations aimed at stemming the short-term negative effects of AI, he warned that it was “difficult” to suggest regulations aimed at limiting its development.
Azeem Azhar, founder of Exponential View, also highlighted the need for a much stronger institutional response “to ensure that AI gives people superpowers rather than being seen as a tool to replace them.”
Mme Bogdan-Martin welcomed the fact that governments and other actors have recently “rushed to establish protections” and regulations around the use of AI.
For example, the European Union announced on Wednesday the creation of an AI office to regulate artificial intelligence under a new law.
