Debate at the World AI Cannes Festival: Can AI correct the ills of social media?

"Right now, the goal is engagement and money and that's it."

A collection of the globe’s most prominent ambassadors of artificial intelligence sipped wine and networked into the night at the home of Franco-American tech entrepreneur Jean-Claude Goldenstein in Cannes, France last week.

Among them were: Francesca Rossi, AI Ethics Global Leader at IBM; Jean-Paul Paoli, who runs generative AI for beauty giant L’Oreal Group, Eric Garandeau, TikTok’s director of public policy and government relationships; Adam Cheyer, VP AI Experience at Airbnb and Sebastian Hallensleben, who is part of the EU’s effort to create standards to underpin regulation as chair of CEN-CENELEC JTC 2.

They were all in town for the third annual World AI Cannes Festival (WAICF) created by former Apple chief operating officer, Marco Landi.

Photo: AI intelligentsia gather to talk about AI for good.

Photo: Creopoint’s Jean Claude Goldenstein, expert on tracking disinfo and deepfakes using both people and AI.

First the fun: I met “Captcha,” (below) whose schtick is that he’s sarcastic and puts down humans for asking banal questions. You can see him here. Hidoba Research reps told me they are hoping to find partners like hotels to put the robot to commercial use. The company’s business cards warn that they may use AI to answer phone inquiries however. There was also a hologram which might one day help you locate the gate at the airport. The Tesla robot Optimus shook hands and the French Air Force let attendees do flight simulation.

Byte Me: Captcha, the rude robot.

All fun and games on the show floor but the lecture sessions were much more serious.

AI and Social Media.

In one panel on AI ethics, Fred Werner, head of the strategic engagement division at the United Nations International Telecommunication Union (ITU) hoped that AI could be used to correct the mistakes of social media by delivering positive body images to youngsters.

“If somehow there was more money to be made from that, than just optimizing for engagement, then you’d have a very different scenario but the incentives are just not there and it’s the same for presenting truth or misinformation. The technology can be used for good or bad.” 

Werner doesn’t think regulation will change much without financial incentives.  “You can provide guidelines and encourage behavior with taxation,” but he wasn’t optimistic. “Right now, the goal is engagement and money and that’s it.” 

Sorbonne University Professor Jean-Gabriel Ganascia pointed out how hard it is to create rules that work for every kind of society, while America favors individual enterprise - the Chinese do not. Who could have foreseen some of the unintended consequences of social media when Facebook was in its infancy? Regulators do not have an easy task.

The ethics panel, which also included Albena Kuyumozhieva, a policy officer at European Innovation Council of the EU, touched on some serious threats that could makes social media even worse. She pointed to examples of AI generated nudes of young girls that were shared by classmates in Spain. El Pais’ story referenced the app Clothoff, which carries the slogan, “Undress Girl for Free.” 

The Future of Life Institute wants to slow AI development. It’s home page states that, “more than half of AI experts believe there is a one in ten chance this technology will cause our extinction.” They fear AI is akin to nuclear weapons and potentially on the way to destroying us all, an analogy not embraced by IBM’s Rossi or Meta’s Chief AI scientist Yann LeCun when asked about it in another session.

LeCun says it will be a very long time before AI becomes more intelligent than humans. “AI is going to amplify human intelligence,” he said, “But we shouldn’t feel threatened by this.” If people suggest AI will end humanity,  “don’t listen to them.” More on what he had to say here.

A presentation by Shutterstock’s senior director of AI and data science, Alessandra Sala, showed the dangers of race bias in AI in a slide illustrating doctors and lawyers as older white men and people of color in photos of garbage collectors. She explained how Shutterstock puts a premium on supporting human creators while also offering up a tool to create AI-generated images that are clearly labeled.

Meta too will detect and label more content, AI-generated, in the run-up to the Presidential election in the US.

At coffee meetings and in sessions, conversations were mostly about how to keep AI in the realm of doing good, with open source development a priority. The European Union isn’t going to let a handful of American companies dominate the space by erecting their platforms on top of the latest oil gushers of technology. It is actively supporting new AI development with research funds worth 290 million Euros.

Good Luck with that. OpenAI’s Sam Altman is seeking up to $7 trillion for more chips to supercharge AI development, potentially with help from the UAE, according to the WSJ.

The EU is set to rubber stamp its AI Act in April, the first AI legislation of its kind with an agency to impose huge fines on transgressors too. It includes such things as the rights of consumers to have their complaints addressed. If you find your book or your art has been plagiarized who would you call? The EU is taking seriously the threats posed to fact-based reporting as more people find their opinions formed by videos shared without context (or with incorrect context) on TikTok and X. More on the proposals in the act here.

I texted a friend about the AI event in Cannes and it turned out they were also attending an AI conference in their part of the globe. AI gabfests are a mushrooming business in every field. The business world is desperately trying to keep up with advancements, but the promise of the AI revolution feels less well articulated then its threats.