Fighting online abuse using AI
With athletes noting increases in cyberstalking and harassment, Notre Dame is leveraging the power of AI to identify and stop abuse.
Excerpt from The Players’ Tribune, written by Lily Fenton, originally published on March 5, 2026.
“When the terror finally comes, it’s at night.
It’s a Thursday, maybe like 1 a.m., and you’re the only one home. Two of your roommates are out, the third is away at a study session. And you’re on your phone, setting your alarm before bed. A text pops up. But this time the text is different. This time, at the end, there’s not just a link—there’s an actual picture attached. It’s a picture that was originally from social media, of you with your friends and family. Now, though, it’s been edited with AI.
Edited so you’re completely naked.
Your stomach drops. Your whole body starts sweating and shaking. Now you know the answer to one of your questions: This is really happening. That picture does not exist . . . only it does exist. You’re looking at it right now. You’ve never felt so powerless.”
Lily Fenton is a senior at Notre Dame and a setter on the women’s volleyball team. But for the early part of her college career, she was worried about more than how to balance practices with her intensive engineering coursework. Fenton had an online abuser who, for months, flooded her phone with text messages, phone calls, and eventually, inappropriate and offensive photos. Fenton recalls feeling ashamed, confused, and alone throughout the experience.
But she is far from alone.
Stats on online athlete abuse
Monitoring from various groups shows online abuse of athletes is a growing problem.
- 7,085 abusive, discriminatory, or threatening messages verified in the 2023 FIFA Women’s World Cup. (FIFA)
- 5,000+ of 1.3 million posts and comments directed at student athletes were determined to contain abuse. (NCAA)
- 51% of D1 men’s basketball players received social media abuse related to athletic performance. (NCAA)
- 46% of negative or threatening messages were from someone who bet on a game. (NCAA)
- 49% of flagged posts from the 2024 Paris Olympics were directed at U.S. athletes. (World Athletics)
- 50% of abuse directed at athletes is misogynistic or racist. (United Against Online Abuse)
Why athletes face abuse online
FIFA verified 7,085 abusive, discriminatory, or threatening messages during the 2023 Women’s World Cup.
During the 2024 Paris Olympics, World Athletics monitored 1,917 track and field athletes. Of the 355,873 posts it analyzed, 809 were flagged as abusive. Of those highlighted posts, 49 percent were directed at US athletes.
And abuse is spreading from professional sports into college athletics.
In recent years, the NCAA has released several studies on abuse in athletes. The 2025 NCAA GOALS study of 20,000 student-athletes revealed that 51 percent of Division I men’s basketball players received social media abuse related to athletic performance.
The study also considered why athletes were being abused. The 2025 survey found 46 percent of negative or threatening messages were from someone who had bet on the game. A 2024 NCAA study combed through 1.3 million posts and comments directed at student-athletes, and more than 5,000 were determined to contain abuse. Of that abuse, 18 percent was sexualized, 12 percent was related to sports betting, 10 percent was racist, 9 percent was homophobic or transphobic, and 6 percent contained violent content. Much of the study was directed at athletes participating in March Madness, and women’s basketball players received three times as many threats as men’s.
Fenton also mentioned the NCAA name, image, and likeness rules, or NIL. As a member of one of the first eligible classes, Fenton remembered the excitement of posting online and hoping for brand deals and notoriety that would pay off. But because NIL was so new, there wasn’t a lot of consideration of potential downsides.
In August, the NCAA announced it will help the FBI fight cyber-enabled crimes and offer affected athletes adequate support, but more work is needed.
Motorsport leads the race to curb online abuse
Motorsport is also a hotbed of such online abuse in comments on social media, direct messages, or, like in Fenton’s case, to their personal phones. Drivers have endured death threats, xenophobic comments, and vitriol in the wake of losses. Fake photos have cost them brand deals. Some claim it has led to a decline in mental health. Others are considering leaving the sport. And it isn’t just the athletes—referees, volunteers, staff, and journalists are also targets of this abuse.
H.E. Mohammed Ben Sulayem is a former rally driver and the current president of the Fédération Internationale de l’Automobile (FIA), the body that oversees global motorsport including Formula 1. He said that F1 fans are very passionate about individual drivers or teams. Social media gives those fans fun, interactive access to drivers, but fans are increasingly taking things too far when criticism turns into threats.
“FIA reaching out to Notre Dame speaks to what we as a university stand for. Our mission as a force for good, not only in the United States, but globally.”
- Nitesh Chawla, the Lucy Family Director for Data and AI Academic Strategy at the University of Notre Dame
Ben Sulayem shared the story of an FIA steward and race director who received threats to rape and kill her after a controversial penalty at the US Grand Prix in Austin, Texas, in 2022.
More recently, Kimi Antonelli, then a 19-year-old driver for Mercedes, received more than 1,100 “severe or suspect comments,” including death threats, after the 2025 Qatar Grand Prix. Antonelli had been vying for third, but made an error that allowed McLaren driver Lando Norris to whiz past him. Fans accused him of letting Norris through, essentially cheating.
While many of the threats remain online, others can become in-person crises including stalking, doxxing, and more. The severity of abuse online, combined with the safety and mental health of officials, athletes, and volunteers, spurred Ben Sulayem.
“I said, ‘That’s it. We have to take action.’”
In 2023, under the leadership of its president, FIA launched its United Against Online Abuse (UAOA) initiative. Alongside a number of global sporting entities, governments, and partners, UAOA aims to leverage research and awareness to reduce online abuse in the sporting ecosystem. In its first 18 months, UAOA published two barometer reports that studied the extent of online abuse by leveraging collaborations across FIFA, World Tennis, World Rugby, the International Esports Federation, Fédération Internationale de Motocyclisme, and more.
But Ben Sulayem believed more data would lead to even more understanding. So he reached out to Notre Dame.
Joining forces with the FIA to fight online abuse
As FIA’s UAOA was building its network, its leaders were introduced to Nitesh Chawla, the Lucy Family Director for Data and AI Academic Strategy and the Frank M. Freimann Professor of Computer Science and Engineering at the University of Notre Dame. A leading voice in artificial intelligence and data science, Chawla has made foundational contributions to AI while applying his research to challenges in health, education, scientific discovery, and public services.
“My career has absolutely focused on innovating AI algorithms and technology that have a translational impact on society. We do amazing fundamental research that gets published in the top-tier publication venues, but we also always ask ourselves those difficult questions: How has it made a difference to someone’s life, a community and an organization?” he said. “That’s what keeps me grounded at Our Lady’s University. That’s what our mission is.”
“It’s our core mission that when we take on research, we are not simply taking on research for the fundamental questions that we aspire to answer. But at the end, we ask ourselves what impact can we demonstrate as a public good? What impact can we demonstrate as a societal good?”
- Nitesh Chawla
According to Chawla, UAOA had heard of Notre Dame’s reputation for boldly tackling societal challenges using data and AI, so they approached Chawla about leveraging his expertise. He saw a natural alignment. Both groups had a shared interest in work that was research-led, impact-driven, and global. Together, they intend to build AI algorithms that identify and stop online abuse against athletes.
“When the FIA shared its efforts to combat online abuse of athletes, the need was clear. This is a problem that affects everyone from young student-athletes to professionals. We asked how we could bring our skills, technology, and ideas to help tackle it together,” Chawla said.
As the only US university partner in the coalition, Notre Dame’s role was first to design a survey for American student-athletes. To do so, Chawla’s lab is working closely with Notre Dame’s Athletics Department.
Those survey results, along with data culled from across social media, will help inform the current definition of online abuse. This summer, Chawla hired a team of students who flagged tens of thousands of posts on social media platforms for abusive content so they could gather training data for the AI algorithms to learn about abusive messages.
“We are curating a lot of data, collecting a lot of data from Reddit and other social media platforms to understand . . . what does hate speech or abuse look like?” he said. “Because often it’s camouflaged.”
Training an AI algorithm to pick up on not only flagrant content, but also nuanced, hidden messages, is a challenge. So is anticipating where cruelty is headed.
“We want to develop these AI models to identify:
- Is the trend of abuse changing?
- Are there new signals emerging?
- Are there new sentiments emerging?
- Are there new slang words emerging?
- And is the quantity increasing?
- And is the quantity simply increasing because someone is using AI out there to do whatever they want?”
Once the platform is a reliable indicator for abuse, then comes the hard part: stopping it. To do so, Chawla is in conversation with Signify, an ethical data science company whose AI-driven Threat Matrix has been employed by the NCAA and the International Olympic Committee to fight online abuse.
“What we are developing is an artificial intelligence platform where we can quickly catch, get alerted to, detect what the abuse may be, and then coordinate with the social media platforms as best as we can,” he said. Eventually, he hopes that athletes will be able to subscribe their accounts to a service that will identify and filter harmful posts, before they ever see them.
That is the caveat here—Chawla’s work can flag these comments, but he cannot stop bad actors. Instead, he hopes his data allows law enforcement, policymakers, and governments to act.
“We may not stop the origin of these things, but the question is, can we identify them early enough, flag them early enough such that the systems and the organizations and the authorities can take action early enough, fast enough?”
Ben Sulayem echoed that increased awareness and accountability are key to stemming the tide of abuse. He hopes the data Chawla and other partners gather help them better understand perpetrators and create evidence-based solutions to deter them. In some cases, that means legal prosecution, and FIA and UAOA hope to see more of that.
“Once there is the awareness paralleled by a punishment, I feel that we will be on the right track,” Ben Sulayem said. “We can bring [online abuse] down and we can make sure that these people are [held accountable] for anything or misbehavior.”
Moving forward after online abuse
Fenton’s situation has an unusual ending: Her perpetrator was caught and faced legal consequences for his actions. Most online abuse doesn’t have resolution or justice. Still, Fenton grieves the time the situation took away from her.
“This situation took away almost two years, half of my college career,” she said. “I was dealing with the case, the person, and the situation, but also just the effects of how that changed me as an athlete and me as a person.”
Part of that change meant reclaiming her power, which led to The Players’ Tribune article.
“I decided to write The Players’ Tribune article because I felt like if I had any platform at all, I wanted to use it to reach out to those who have been affected, or could be affected, and just spread awareness to what the things that could happen, the things that have happened to me and to others,” she said. She also wanted to show that there is life after abuse, and there is light after darkness. “There are good people out there that fight for justice.”
Many of those good people have been with her at Notre Dame. Fenton credits her coach, her teammates, the Notre Dame Police Department, and now Chawla’s team for supporting her and trying to create a safer future for athletes like her.
“I think about all the different ways in which Notre Dame specifically has not only helped me in my process, but has shown commitment to the issue as a whole,” she said. “It makes me feel so extremely grateful that I’m part of a bigger family and a bigger community that is taking on these issues, and so I’m not alone in doing it.”
Fenton and Chawla have since connected on campus. He recalled being taken by her courage and, as a dad of teenagers, feeling enormous empathy for the situation she endured.
“They are students to us as much as they’re athletes to us,” he said. “[This project] resonates with the University’s mission on research for good, and how we are looking at AI as a technology that can do harm, but at the same time, we can use AI as a technology that can advance the dignity of a human person.”
While Fenton initially had some reservations about AI, Chawla’s group has been able to show her its benefits too.
She said, “I think for a long time after what I went through, I was very skeptical and reserved about the use of AI and the advancement of AI. But in my conversations with the group and everything, I think it really goes to show that at the end of the day, it is a tool, and all tools can be used for good and bad. And so it’s heartwarming to see it being used for good.”
A moral future: AI ethics and education
Since its 2023 launch, United Against Online Abuse has gathered more than 80 partners, including governments, sports federations, and tech companies, and the invitation is open to more.
“We are asking more partners to come with us. Partners are not only federations of sport . . . but partners come from the technology side, from the governmental side, from the teams themselves,” Ben Sulayem said. “Is it something that we can do alone? No. It is a marathon, and unless we are together, we will fail.”
Chawla’s Notre Dame team is also looking to the future, developing educational programs for students from elementary through high school.
“Students may encounter online abuse in different ways. They may be targets, they may witness it, or they may engage in it themselves,” Chawla said. “How can we help them understand what is acceptable, where boundaries are crossed, and how their actions affect others?”
The goal is to help young people become more aware of their online behavior and the consequences of misusing AI.
Chawla also draws inspiration from Pope Leo XIV’s Magnifica humanitas, the encyclical on safeguarding the human person in the time of artificial intelligence.
“Magnifica humanitas gives us a moral charge and a responsibility to think about how we respond to a world shaped by AI,” he said.
That responsibility guides Notre Dame’s RISE AI, a framework for advancing AI that is responsible, inclusive, safe, and empowering.
“Our work with the FIA is an example of what RISE AI means in practice: bringing together our research, technology, and partners to address a problem that affects people’s lives,” Chawla said.
The What Would You Fight For? campaign is supported in part by a generous gift from David and Debra Delahanty.