ENGLISH TEXTS
� PAGE 64
EQUIDIA
“ THERE’ S A BE- FORE AND AF- TER CHATGPT”
By Serge Okey
One can hardly imagine all that goes on behind the thousands of hours of broadcasts and the vast amount of information shared on Equidia. Behind the scenes, teams juggle daily with algorithms and artificial intelligence. When it comes to tech, the European leader in live coverage has already shifted into second gear- and third is approaching fast, as many projects are already in the saddle. We caught up with David Deschamps, Chief Technology Officer( CTO) at Equidia, for an update.
Galorama. Let’ s start with Enia, the famous AI mascot at Equidia who rates the runners in the Quinté + using artificial intelligence?
David Deschamps. Truth is, it’ s not entirely AI. We use algorithms based on many criteria, but those are weighted by our experts. It’ s mainly data-driven: the horse’ s performance, how well it turns right or left, the distance, the ground condition … Then we add a human touch: our journalists call the trainers, and their insights are factored in. Alongside that, we are working on fully AIbased algorithms with no human input, but the results are not yet convincing. Our journalists are better. We’ re currently drafting a charter to frame all this properly.
G. Do you have all the data you need?
D. D. What we’ re missing is data on foreign horses, debut runners, and training sessions. We tried putting sensors on horses during training, but trainers aren’ t quite ready for that yet, which we understand.
G. Is the use of digital tools mostly limited to betting predictions?
D. D. Far from it. We’ re developing a new chatbot for races beyond the Quinté +. We also use image recognition tools. AWS Recognition is incredibly powerful! Recently, we’ ve started taking screenshots of each horse at the finish line to use in promoting upcoming races. That has real value. We’ re discussing with France Galop the idea of offering owners a photo of their horse at the finish along with a brief summary.
G. Are avatars part of your plans?
D. D. It’ s already being done in Switzerland, where a digital human presents the weather forecast. We’ ve tried it— and will probably do it again. But our charter is clear: only journalists appear on air. With this technology, anyone can be modeled using a webcam. With the rights holders’ permission, we could even“ resurrect” Léon Zitrone. It’ s very trendy on social media. We’ re also exploring this idea with the PMU for their kiosks: using an avatar to guide bettors.
G. Is tracking part of your daily routine?
D. D. We’ re working on it extensively with McLloyd. It’ s an incredible source of data. Two years ago, we had a project to add small tags above the horses during live races so viewers could identify them easily. Hong Kong has done it. We’ re also testing more advanced cameras to track horses’ positions during the race.
G. Are you using ChatGPT?
D. D. Yes, internally— but not yet in production. More precisely, we’ re using Claude, a competing and more efficient algorithm, for writing horse biographies. It’ s still in testing. We need to fine-tune it. All these tools need training. But there’ s definitely a before and after ChatGPT— it changed everything.
G. Are these technologies already impressive?
D. D. Absolutely. The world of technology is evolving at lightning speed. Amazon’ s tool is incredibly efficient with both text and image. It can detect a horse’ s acceleration. Like a Tesla, it remaps the entire environment. Every single object is identified.
151