Dressage riders and officials convened in Ermelo, the Netherlands, last week for the annual meeting of the International Dressage Trainers Club, International Dressage Riders Club, Dressage Organizers Club and the Dressage Officials Club. 

Members of the press were not granted access to the meeting, but Linda Keenan, secretary general of the International Dressage Trainer’s Club, and Klaus Roeser, secretary general of the International Dressage Riders Club, hosted a press conference afterward. 

Rebuilding public trust in dressage sport is a top priority for all stakeholders at the meeting, they said. 

“There was an agreement that we do indeed have significant issues about this sport, which center around perception, welfare, transparency and trust, and everybody agreed that we have collective responsibility to address this and really see what we can do to improve the situation,” Keenan said.

Revamping judging—and equipping judges with machine-learning tools—was another major topic of the meeting, and it was, for the most part, embraced by attendees as a path forward. Though it sounds like something out of a futuristic science-fiction novel, technology-assisted judging is on the horizon. 

This year, research physicist David Stickland launched a startup company, Global Equestrian Technology, which is developing a software that measures horses’ movement using modern video analysis tools. It can precisely track skeletal points on a horse’s body and use them to measure things from whether the horse executed the correct number of one-tempi changes in a test, to how far it moved forward in its 12 best steps of piaffe, to more subtle things like the angle of its face relative to its chest or how much it over-tracked in the walk. The program can be trained to correlate these so-called “observables” with numerical scores typically assigned by judges.  

Although casually referred to as artificial intelligence, Stickland emphasized that he and his team prefer the term machine learning, “as AI suggested a form of intelligence,” he wrote in an email to the Chronicle.

“We want to emphasize that these systems only learn from us and don’t create an independent meaning about our sport,” he wrote. “It is absolutely not our intention to replace judges. It is, however, to assist judges by providing additional information, e.g. biomechanical movements analysis information. 

“We think this will allow judges to better focus on things we do not (yet) see with our biomechanical analysis,” he added. “Whether in the course of time our system will be so advanced that it can judge as well or better than human judges remains to be seen, but technological developments will continue at a rapid pace and may one day reach this stage.”

Keenan said the dressage groups that met last week welcome the idea of incorporating such technology to help relieve human judges of some of the instantaneous work expected of them during a test.

“I think from a rider and trainer standpoint, we are very much supportive of using AI or machine assistance—whatever you want to call it—to help with the judging, in one sense, to a little bit relieve the task of a judge,” she said. “Because it’s such an incredible task, judging. So, this would let them focus on maybe some of the more ephemeral aspects, if you will, about the performance.” 

Dressage judging, an endeavor that is inherently subjective, could still be benefited from objective analysis provided by AI tools, Roeser noted. 

“There are a lot of possibilities, especially when it comes to the technical part, like counting and measuring,” he said. “For example, tracking traveling in the piaffe, the circle of the pirouette, counting the amount of tempis, et cetera. Those are easy to track.” 

Riders and trainers would want the assistive technology tested first, Keenan said, to mitigate unintended consequences. Judges have their own concerns. 

“The judges are a little bit scared about it,” Roeser said. “I think they may think, ‘OK, maybe in five years, they don’t need us anymore.’ But we told them that that’s totally not the point. The complexity of judging is so high, so we would like to reduce that. Then, the judges can even more concentrate on quality, because quality you can’t measure with AI.” 

The more subjective aspects of dressage, such as the quality of a horse’s movement, will not be measured by computerized tools, in part because of the sheer volume of data that would be necessary to train such an advanced program. Even if the quantity problem was surmountable, modern dressage horses, thanks to evolving breeding standards, have changed so rapidly in recent years that AI tools would not be able to adapt fast enough, Roeser said. 

Stickland does, however, believe technology has the potential to reshape the way the sport of dressage is judged.

“Here we are absolutely convinced,” he wrote. “Machine learning is making its appearance in all areas of society, so also in sports. When our system would be used, we can reduce the subjective bias that is undeniable and understandably present among judges.  

“We will also be able to reduce the inevitable ‘misses’ of judges,” he added. “By this we do not mean to criticize current judges, but that the task given to judges is in reality impracticable. At the time of the FEI [Dressage Judges Working Group] review, an independent group of specialist psychologists advised it would be almost impossible for a human brain to execute visual cognitive decision-making consistently under the conditions, and for the duration, the FEI currently expect them to do. Inconsistent decision-making can lead to unjustifiable and unfair scoring and is one factor that can result in a loss of integrity to the sport.”

Spectators, too, could benefit from the introduction of assistive technology to dressage judging, Keenan predicted. 

“It could help them understand the scoring a little bit more,” she said. 

In practice, technology-assisted judging could take many forms. Judges could opt to use only the tracking technology, to help them determine whether the combination executed the movements accurately. Alternatively, judges could use the technology to count things like steps in a reinback, or the number of flying changes, and issue a pop-up screen to notify them when a horse performed too few or too many. Nothing is set in stone, as testing has yet to begin in official competition settings. 

“There’s an agreement with the FEI that we are going to test some possibilities in the very near future,” Roeser said, starting with the simpler tracking abilities to see if those are helpful for judges.

Global Equestrian Technology’s work is still in the development stage, but Stickland expects to have his software in judges’ boxes soon. 

“When and how we will be able to use it in competitions is dependent on how well this development will progress and how much budget we will be able to free up,” he wrote. “We don’t want to make … hard predictions, but we have to think in terms of years rather than months. This is not to say that we cannot start the first test earlier.  Our ultimate goal is to be able to deploy it in [the 2028 Los Angeles Olympics]. Perhaps even more importantly we plan to progressively make it available to test in competitions and for individual riders to use it at home between now and LA.”

Addressing Welfare Remains Paramount

Horse welfare, including whether to allow snaffle bridles in Grand Prix competition, was also a topic of discussion at the annual meeting. 

In years past, the IDRC and IDTC have stood firmly against snaffle bridles in Grand Prix competition, but the group’s stance appears to be softening slightly as national equestrian federations, and the FEI, have begun to open the door for snaffles in Grand Prix classes. 

Still, Keenan emphasized, the question of whether to allow snaffle bridles is unquestionably not a welfare issue.

“Riding in the double bridle is one of the ultimate skills of dressage riding, and it’s something that needs to be tested,” she said. “Competitors need to prove their competence at the top level.”

Earlier this year, the Swedish and Dutch dressage federations made a proposal to stop requiring double bridles in CDI3*/CDIO3* competitions. Although the proposal was initially met with skepticism, the FEI’s technical committee has since suggested running experimental Grand Prix classes to be ridden in a snaffle bit next year. 

“Many of the national federations permit a snaffle bridle in Grand Prix, and we don’t have any problem with that,” Keenan said. “And we don’t have a problem, in principle, in FEI competitions using the snaffle bridle if those classes would then not be Olympic qualifiers.

“But if there’s a desire to have Grand Prix classes in a snaffle, perhaps with younger horses or whatnot, we could see that it could have a place. We’re happy to see what happens,” she continued. “Let’s see what the appetite is. But our concern is that the double must remain for Olympic qualifying classes.”  

Newly appointed FEI Director of Dressage Ronan Murphy is in the process of developing a comprehensive strategy intended to address all the challenges dressage sport faces today, including regaining the confidence of the public. Though neither Keenan nor Roeser knew when to expect the unveiling of Murphy’s plan, they remained optimistic. 

“I can’t say too much, because I don’t want to steal Ronan’s thunder, but we are going to be doing specific things to gain more trust, and we will be actively promoting and participating in those things,” Keenan said.