The system will analyze public comments and social media conversations to identify potential risks, alerting the FEI's communication and security teams in real time. It will also assist in moderating official event channels, managing private messages, and providing support for those targeted by harmful content.
Growing Problem of Online Abuse
The decision to deploy this system is a response to real and alarming incidents in the equestrian community. Over the past few years, riders, officials, and even their families have been subjected to extreme online abuse. This includes death threats, invasive personal attacks, and the public release of private information such as addresses and workplace details. These threats do not exclusively target world-famous athletes but also extend to judges, stewards, and volunteers. For example, a judge’s decision in a dressage competition can lead to their personal details being shared in online groups masquerading as animal welfare groups, even though those claims may have no factual basis.
One major problem is the lack of institutional support for those who receive such abuse. Victims have had to manage these situations on their own, collecting evidence through screenshots and seeking legal help to address the harassment. The AI monitoring system is intended to shift this burden from the individual to the organization by providing structured documentation and evidence management for online threats. This is especially important given that content is often quickly removed, and accounts are taken down, making it difficult to preserve any proof.
Concerns and Limitations of the System
However, despite the clear need for online protection, the current framework raises significant concerns. The FEI is not only the entity implementing the system but also the main target of much of the critical commentary. This creates a potential conflict of interest, as the organization is both the arbiter and the subject of the content being monitored. The lack of independent oversight means the FEI retains the power to decide what qualifies as harmful content, which could lead to the suppression of legitimate criticism. The system combines two distinct functions: threat detection and reputation management. This overlap could blur the lines between safeguarding individuals and managing the FEI’s public image, raising concerns about misuse of the system for censorship rather than protection.
Data privacy is another area of uncertainty. This involves handling personal information from EU-based users to generate risk assessments. Yet, it remains unclear who is legally responsible under GDPR – is it the FEI, the FN, or Signify itself? What legal framework authorizes the creation of user profiles? How long will the collected data be retained, especially if the content turns out to be harmless? These questions remain unanswered, leaving potential privacy violations unaddressed.
The system also has technical limitations. It struggles to interpret content written in regional dialects, ironic statements, or specialized equestrian language that may not be genuinely threatening. For instance, a discussion about animal welfare practices could be mistakenly flagged as hate speech if the system lacks the context to distinguish between constructive criticism and harmful intent. The error rate of such automated systems is not disclosed, making it difficult to assess the accuracy of the monitoring. This creates the risk of false accusations and the marginalization of legitimate concerns, particularly in highly polarizing topics like animal welfare in equestrian sports.
Testing and Contextual Considerations
Despite these issues, the system has been tested in previous events, including the World Cup Final in Fort Worth. The language of many attacks is in German, and local officials have a better understanding of the country's legal landscape and the cultural dynamics of the sport. While the pilot has shown potential benefits, the integration of the system into the FEI’s operations at Aachen remains an ongoing process.
In sum, the use of AI for online monitoring in equestrian sports is a step toward addressing a pressing problem. However, the system’s structure invites questions about transparency, independence, and fairness. It highlights the need for clear boundaries between protection and control, ensuring that the intent to safeguard individuals does not result in the suppression of legitimate voices.