More than 85% of men’s professional tennis players who received engagement on social media encountered online abuse during the past year, according to new figures showing the growing scale of automated moderation in professional sport.
Reuters reported that the ATP’s Safe Sport programme analysed more than 3.7 million comments across player and ATP social-media channels during the period. More than 284,000 were identified as abusive.
The system prevented around 234,000 abusive comments from reaching players or fans, with more than 82% of detected abusive messages removed in real time.
Tennis is turning moderation into safety infrastructure
Safe Sport combines automated detection with specialist human threat assessment rather than relying entirely on content filters.
ATP originally developed the service with Tennis Data Innovations, Sportradar and Arwen AI. Its initial system was designed to scan social-media channels across 29 languages and automatically hide toxic or abusive comments, while investigators handled more serious threats.
That model is different from ordinary platform moderation.
Instead of waiting for Instagram, X, TikTok or another network to decide whether a comment violates its rules, sports organizations can operate an additional safety layer around the people most likely to be targeted.
ATP Director of Safeguarding Andrew Azzopardi said preventing more than 234,000 comments from reaching players demonstrated the programme’s practical effect, adding that the organization would continue investing in technology, expertise and partnerships.
A small number of accounts can generate enormous harm
Evidence from the wider tennis ecosystem shows why simply counting abusive posts can obscure how attacks happen.
A July report from the ITF and WTA found that their separate Threat Matrix system analyzed more than 1.2 million posts and comments during 2025. More than 12,000 were classified as abusive, including 3,726 containing serious abuse such as violent, sexual, racist or targeted threats.
The same analysis found that just 9% of abusive accounts generated 87% of all abuse categorized as high concern. Angry gamblers accounted for 42% of verified abuse and 59% of serious abuse.
That concentration changes the moderation problem. Identifying persistent offenders can sometimes matter more than simply deleting individual messages.
The problem continues to reach individual athletes
The statistics have visible human consequences.
Earlier in October, Coco Gauff said she received an insane amount of racist abuse during the China Open, describing the experience as draining.
For platforms and product teams, tennis offers an instructive model.
Online-safety systems are increasingly combining automated classification, real-time intervention, human investigation and escalation to law enforcement instead of treating moderation as one isolated technical feature.
The technology cannot eliminate abuse at its source.
But it can change an important part of the experience: whether the target has to absorb the abuse before somebody decides to remove it.
