Thomas Bach has been repeating it like a refrain since last year, in private as well as in public: the Olympic movement cannot ignore artificial intelligence. It will play a major role in its future. Still vague, certainly, but decisive.
The IOC president most often speaks of it in the future tense, as a reality to be grasped without delay. Sebastian Coe, who aims to succeed him next year in the vast presidential office at Olympic House, can speak of it in the present tense.
World Athletics has released the findings and conclusions of a four-year study conducted at the two major athletics events – the Olympic Games and the World Championships – into abuse and harassment suffered by athletes on social media.
Conducted with the help of a cutting-edge tool, the Threat Matrix, piloted by the Signify Group company, the investigation is the first carried out by an international federation over such a long period. It covers the Tokyo 2020 and Paris 2024 Games, and the 2022 world championships in Eugene and 2023 in Budapest.
What should we remember? A handful of figures, first. Massive, for the most part, like the activity on the digital platforms of the best athletes in the world.
In total, World Athletics and its tool, Threat Mix, analyzed over 1,4 million messages or comments posted in around forty languages on social networks across the four events. They potentially concerned the 2.438 athletes participating in the World Athletics Championships and/or the Olympic Games.
At the Paris 2024 Games, 1.917 athletes and officials with at least one active account were monitored, almost 12 times more than three years earlier at the Tokyo Games. In Japan, the study focused only on Twitter. In France, it was also extended to Instagram, Facebook and TikTok. At Tokyo 2020, 240.707 posts were captured and analyzed. At Paris 2024, Threat Mix processed 355.873.
So much for the numbers. Now for the conclusions. They are not short of surprises. The first: in four years of analysis, World Athletics discovered that online abuse was widespread, which was expected, but that it targeted a small number of athletes, which was less so.
At the Paris 2024 Games, of the more than 350.000 messages captured by AI on X, Instagram, Facebook and TikTok, 809 were found to be abusive. But the vast majority of them (82%) concerned only two athletes.
The same finding was made for the Tokyo 2020 Games, where only 132 posts were deemed abusive, out of the 240.707 analyzed. But more than half of them (63%) targeted two athletes, both female.
Another finding of the survey is that hate and insults do not diminish over time. World Athletics explains that three of the five most harassed athletes on social media at the 2023 World Championships in Budapest are still, a year later, in the top 5 established by Threat Mix at the Paris 2024 Games.
The world athletics body emphasizes that its survey is not just for statistical purposes. Its main aim is to protect athletes. "Athlete wellbeing is at the top of our priority list, and we will continue to put measures in place to ensure they can engage confidently and safely on social media platforms" insists Sebastian Coe.
How? Over the four years of its study, World Athletics reported more than a thousand abusive messages (1.258) to the relevant platforms. It identified and denounced 254 accounts that were the source of abusive or threatening content.
Above all, the body offered its help and resources to the athletes most targeted on social networks. Twenty-five of them benefited, throughout the Olympic year, from protection by artificial intelligence. It will be extended over the coming years.
Finally, two cases deemed "serious" were brought to the attention of the authorities, after identification of the accounts and collection of evidence.
