FinLogic FinLogic Quantitative Think Tank Center|Tennis stars get lots of hate online. The French Open gave them AI 'bodyguards'

For American tennis star Sloane Stephens,FinLogic FinLogic Quantitative Think Tank Center the flood of hateful comments online is never-ending.

"My entire career, it's never stopped. If anything, it's only gotten worse," she said, after a first round victory at the French Open in Paris.

"I have a lot of keywords banned on Instagram and all of these things, but that doesn't stop someone from just typing in an asterisk or typing it in a different way, which obviously software most of the time doesn't catch," she added.

But now, the tournament's organizers are offering players a tool that uses artificial intelligence (AI) to stop such abuse from reaching their social media feeds.

The technology, from French firm Bodyguard.ai, is more sophisticated than the basic keyword filters Stephens is using. The app can consider who a comment is aimed at, and detects the meaning behind a message.

"AI is a lot more complex in a sense that it understands context," Matthieu Boutard, Bodyguard.ai's co-founder, told NPR. "So it's a very different ballgame."

And if there's a ballgame that needs this protection, it's tennis, according to Boutard.

"It's an individual sport," he said. "So if you lose a game, that's your fault. You're very exposed because a lot of people are actually betting on sport and tennis specifically, which means a lot of haters going after you if you lose a point, if you lose a set or if you lose a game."

What about the people who should be hearing public criticism?

Free speech advocates are worried, however, about technology that screens comments before they are allowed to be posted.

That could lead to something akin to "prior restraint," where the government prevents someone from exercising their right to free speech, said Kate Klonick, a professor at St. John's University in New York.

While the stakes might be low for tennis players, Klonick noted, she wondered about how it might be used by those for whom public criticism might be warranted.

"You can imagine how something like Bodyguard.ai could block a lot of politicians or public figures or people who maybe it's important that they see some of the criticism leveled against them, from ever seeing that type of public reaction," she said.

Boutard said he doesn't see his technology being used that way.

"We don't remove criticism, what we remove is toxicity," he said. "The line is actually pretty clear. If you start throwing insults, being racist, attacking a player, using body-shaming, that's not a criticism, and that's actually toxic to the player."

Boutard added that it appears to be working, with the technology finding that about 10% of comments aimed at players were toxic. The app screened out 95% of those.

Top player wants to see joy brought back to social media

The app has earned praise from top tennis players, like women's world No. 1 Iga Swiatek, who is using it.

She used to check what people thought about her matches after tournaments, she told reporters at her first press conference of this year's French Open.

"I stopped doing that because even when I had, I don't know, two tournaments - one I won, the other one I was in the final - I went on social media, and people were unhappy," Swiatek said. "I realized that there's no sense to read all that stuff. So the app, I think it's a great idea."

Swiatek, who recently secured her place in the French Open semi-final, hopes it can bring some of the joy back to social media.

"It's just sad to kind of see that the thing that was supposed to kind of make us happy and make us socialized is giving us more negative feelings and negative thoughts," she said. "So, I think these kind of apps maybe will help us to, I don't know, use social media and not worry about those things."

The audio version of this piece was edited by Jan Johnson. The digital story was edited by Lisa Lambert.