“Say, ChatGPT, what impact could artificial intelligence have by 2050 on judo, both as a competitive sport and in terms of the organization of its daily practice in clubs and federations?” Asked to the conversational agent emblematic of the emergence, at the end of 2022, of generative artificial intelligence, the prompt obtains in four seconds an answer in seven synthetic paragraphs, concluding on the high potential for transforming judo into a “more efficient, accessible and equitable” practice.

The seven entries found are: “personalized training“; “opponent analysis“; “safety and injury prevention“; “automated refereeing“; “club and competition management“; “accessibility and inclusion“; “community and engagement“. These seven entries draw a full spectrum encompassing multiple tools: sensors, cameras, biometric data, algorithms, virtual coaching, community platforms… So many terms that, since the eruption in autumn 2022 of OpenAI’s Chat Generative Pre-trained Transformer (GPT), are becoming more and more present in everyday globish vocabulary, in the same way as those whose occurrence on search engines has exploded in recent years: deep learning, machine learning, big data, generative AI, etc. “Innovations are appearing at such a high rate that we’d have to take stock of them every six months,” announces Vivien Brunel, a third-dan polytechnician and doctor of science, in one of his columns for the French bimonthly L’Esprit du judo, back in January 2024.

What is AI? The term AI, popularized in 2001 by Steven Spielberg’s film of the same name, is on everyone’s lips today among that part of humanity that has access to digital technology – according to a report by the International Telecommunication Union, the UN’s technology arm, 2.6 billion people, or 33% of the planet’s inhabitants, had no access to the Internet in 2023; conversely, 67% of the world’s population, or 5.4 billion people, are now online.

What exactly do these two terms mean? A comprehensive article by William Audureau in the French daily Le Monde on April 20, 2024, summarizes: “This is a specific multidisciplinary field, at the crossroads of computer science, mathematics, cognitive psychology and linguistics. It aims to reproduce certain human skills, such as reasoning, planning and problem-solving, using computer algorithms. The term ‘artificial intelligence’ was introduced in 1956, in reference to the work of French positivist philosopher Hippolyte Taine. In 1870, in “De l’intelligence”, he compared human intelligence to a ‘machine’ whose mechanism AI pioneers hoped to model and reproduce.” However,” continues Karine Deschinkel, a teacher researcher in computer science at the Université de Franche-Comté, in the same article, “we shouldn’t say intelligence because it’s been programmed. It’s an algorithm, almost like a recipe, a set of mathematical calculations that teaches the machine to carry out tasks normally performed by humans.” The article goes on to list the various fields of application in our daily lives, gradually infused by this sprawling tool with exponential computing power: search engines, predictive polite phrases suggested in our e-mails, targeted advertising, automated translation, chatbots, weather forecasts, telephone call centre animation, profile analysis for recruitment, road traffic prediction, medical image analysis, inventory management, automated writing of journalistic briefs, autonomous cars, smart thermostats, connected objects, industrial robots, the pharmaceutical industry, analysis of medical imaging results, etc. On this last point, Le Monde inserts an enlightening observation by Laure Abensur Vuillaume, a specialist in emergency medicine at the CHR Metz-Thionville: “It’s a kind of blinker, a warning. AI doesn’t do better than the doctor, but the AI-doctor pair does better than the doctor alone.“

Ukraine-Gaza: An Open-Air Laboratory. Regarding the conflicts that are ravaging the planet in this first quarter of the 21st century, AI is also becoming a key issue. In this regard, it’s worth reading the very solid study delivered in September 2024 for the French Institute of International Relations (Ifri) by Amélie Férey and Laure de Roucy-Rochegonde, respectively the coordinator of Ifri’s Defense Research Lab and the head of Ifri’s Geopolitics of Technology Center. Titled From Ukraine to Gaza: Artificial Intelligence at War, the study notably cites +972, an Israeli-Palestinian investigative journal, which recently “described the operation of two other software systems, Lavender and Where’s Daddy?. Lavender assesses the probability that a Gazan belongs to an armed organization by comparing their communication patterns (frequent phone number changes, contacts with numbers affiliated with these organizations, etc.) to those of known members of Hamas or Islamic Jihad. Where’s Daddy?, on the other hand, locates targeted individuals when they return home and alerts officers to their presence in identified buildings so they can fire, despite the presence of civilians.” Scientifically stimulating but humanly chilling prospects, given the inevitable margin of error at this stage of technological advancement and the potential human cost.

Beating Garry Kasparov. What about AI applied to the world of sport, and in particular to the tip of the iceberg that is competitive sport? The meticulous way in which the major nations prepare for each event is bound to make eyes soft at such developments – cf. Jacques Nienaber and Rassie Erasmus’s tour after tour of South Africa’s masterclass at the 2023 Rugby World Cup, where the team’s attention to detail went so far as to prepare for their quarterfinal against the French hosts by placing huge amplifiers at the edge of their training ground to accustom the players to the deafeningly hostile reception at the Stade de France a few days later…  In his aforementioned article, William Audureau recalls that “the American economist Herbert Simon (1916-2001) prophesied that the machine would beat the human at chess before 1967“, but that it finally took another thirty years, until 1997 and “the democratization of far more powerful microprocessors, for an algorithm, Deep Blue, to beat the world-renowned grandmaster Garry Kasparov. ” The truth is the fields of application are considerable. Refereeing, statistical and technical-tactical analysis, planning sessions and seasons, injury prevention… When it comes to finding the small percentage that changes everything, the potential levers are infinite – see the meticulous analytical work of Geir Jordet, a soccer psychology researcher at the Norwegian School of Sports Science, known for his image-by-image deconstruction of players’ body language during penalty shoot-outs to understand the mental dimension and hidden motives.  

Range of possibilities. In the United States, Michael Lewis’ book Moneyball (2003) and its film adaptation The Strategist by Bennett Miller (2011) have done much to give credence to a statistical approach to professional sport – baseball, in this case. Ice hockey and basketball are also ahead of the game in this area, as illustrated by the Toronto Raptors’ NBA title in 2019, built from the 2016 draft on an ambitious and innovative partnership with IBM around data – see also the Stephen Curry effect at the Golden State Warriors, where the Californian point guard’s extraordinary three-point shooting statistics have permanently altered the offensive schemes of an entire generation. Data is also used in France, where President Damien Comolli’s Toulouse Football Club has completely overhauled its recruitment system, relying on this data to promote better decision-making, strategies and scheduling of sessions. Still in soccer, statistical measurement tools are already being inserted into shin guards, while television broadcasts abound with a growing number of virtual animations, expected goals and distances covered measured in real time for each player. In rugby, mouthguards are used to record data. In mountain biking, AI is already creeping into the upstream adjustment of bikes’ front and rear suspensions. In sailing, in the self-steering system and in optimizing trajectories according to weather forecasts. In athletics, it’s possible to set one’s running pace to that of a virtual silhouette projected onto an adjoining wall at the exact speed desired. In each of these sports, connected bracelets are slowly but surely entering the market… Four years after the delicate period of confinement and postponement of the Tokyo Olympic and Paralympic Games, which caused so much damage and awakened so many inner insecurities in top-level athletes unaccustomed to so many uncertainties in their usually meticulous schedule, the range of possibilities extends even to the very intimate field of mental health. In Le Monde dated August 6, 2024, Hélène Pagesy writes, for example, that “on the community site Reddit, several users sing the praises of Therapist and Psychologist, robots that are ‘helpful’, ‘compassionate’ and give ‘the impression of talking to a real therapist, but a good and reasonable one’ […], also available 24 hours a day and free of charge, which can be an advantage when you’re ‘broke’.”

Spectator experience. The revolution is also affecting the area around the pitch and even the stadiums, as illustrated this autumn in France by the controversy surrounding the continued use of algorithmic video surveillance equipment installed during the Olympic and Paralympic Games. In March 2024, sports economist Maxence Franceschi detailed on the X network (formerly Twitter) the pioneering use of Parallel Ads by German soccer club Bayer Leverkusen. “Parallel Ads is a technology that allows ads to be displayed on LEDs at different edges of the pitch, depending on the TV feed. This makes it possible to have a display for the domestic market, a display for rights holders on the Asian market, etc. All this is done at the same time as the image is processed, and without changing the spectators’ stadium experience. This makes it possible to sell these spaces several times over, and at a higher price, by offering more targeted advertising tailored to different markets.” Between July 24 and August 11, 2024, the French daily L’Équipe also innovated in its coverage of the Olympic Games by offering Les JO en 2 minutes chrono, a daily two-minute podcast recapping the day’s program, the medal table and the previous day’s highlights. What made it special? It was “entirely generated by artificial intelligence from L’Équipe content“. As for match fixing, well, they’d better watch out: Steve Austin, the first bionic man played by Lee Majors in The Six Million Dollar Man, or Joi, Ryan Gosling’s holographic girlfriend in Canadian Denis Villeneuve’s Blade Runner 2049, may not be as far away as they seemed in 1973 and 2017, the respective years of the series’ and film’s releases.

The eye of the horse trader. What about judo? As a competitive sport, the discipline has come a long way. Spectators in the stands at the 2018 Paris Grand Slam will remember this moody gesture from Japan’s Hirotaka Okada. A remarkable short-lived referee on the international circuit, the 1992 Olympic medallist and double world champion in 1987 and 1991 in the -78 and -86 kg classes, ostentatiously removed his earpiece during a tense bout in which he was the main referee. With this gesture of defiance and independence, he chose to defer to his own eye, rather than the instructions dictated by the central table. Like the tennis player John McEnroe in his heyday, who justified his legendary fits of anger with an unstoppable “I’ve hit enough balls in my life to know whether they land good or bad“, the Japanese marks the swan song of an old-fashioned judo, where judgment born of practice and experience still takes precedence over the technological screen that has since become the norm. Although… At the 2024 Olympic Games, Frenchman Joan-Benjamin Gaba validated long months of stubborn intuitions on his part from part of the French staff – Baptiste Leroy and Frédérique Jossinet in particular, in coordination with Stéphane Frémont and Richard Melillo’s famous “Special Forces” where Gaba and others had grown accustomed since confinement. Long accused off the record by the entourage of his direct national rivals of being “protected” despite numerous premature eliminations in the early rounds of the competitions in which he was involved, the -73 kg was able to “close mouths” at times when words counted triple: French champion in November, bronze medallist at the European Championships in April, individual silver medallist and decisive in the conquest of mixed team gold at the Olympic Games in July. It’s hard for ChatGPT to predict this sudden ability to perform on the big day. But some of the coaches could sense it.

Pushing the cursor. And yet, in the eternal quarrel that has pitted the Ancients against the Moderns for centuries, it has to be said that AI is already much more than a quivering breeze. In Italy, Emidio Centracchio, older brother of Maria, Olympic medallist in the -63 kg class in Tokyo, has been announcing the launch of his pay platform JudoData in 2023, which takes things a step further than the Dutchman Hans Van Essen’s JudoInside and his little brother JudoBase, the statistical arm of the International Judo Federation, both of which are already concrete on dates, head-to-heads and even news in the case of the former. With the help of AI, the former Italian -81 kg manages to detail and classify the types of techniques used in each bout. In France, former junior -60 kg world champion and double senior -66 kg European medallist David Larose has been a top-level performance analyst with the French Judo Federation since 2022. “With the explosion of AI in all sectors, the possibilities in the field of video analysis are evolving considerably, declares the successor to the young retiree Thierry Loison on the federal website in July 2024. “Artificial intelligence makes it possible to automate part of the analysis, to identify patterns and trends more accurately and quickly, opening up new perspectives for optimizing athlete performance. Video analysis coupled with AI is thus becoming an essential tool for remaining competitive on the international scene.” In Japan, the 2019 world championships provided an opportunity to experiment with 360° slow motion, while three years later, Olympic champions Shohei Ono and Aaron Wolf willingly lent themselves to a 3D experiment conducted as part of Canon’s Volumetric video program, a twirling technology that allows a fight to be filmed from the top of the skull to the underside of the fighters’ feet. A foretaste of what the television broadcasts of the future could be like.

A huge field of action. To take up ChatGPT’s tree-like thinking at the start of this article, there are many judo-specific needs that AI could fill. They can be classified into several families. For competitors, AI could be a plus when it comes to programming, personalizing, and optimizing training, considering their technique, current form, and fighting strategies. At the other end of the spectrum, AI can, when coupled with video analysis as explained by David Larose above, enable a closer study of opponents’ characteristics: right- or left-handedness, high cardio, strong times, weak times, dangerous in standing-to-ground linkups, less vigilant on guard restarts, etc. On the medical front, AI can help prevent injuries by analyzing biometric data and anticipating fitness cycles and slack periods. On the refereeing front – an eternal nagging theme for almost a quarter of a century, with the still-controversial Douillet-Shinohara match in the final of the Sydney Olympics! -there is real room for improvement, for example, in making the distribution of shidos fairer from one mat to the next, or even, who knows, spotting simulated injuries or arranged fights which, one day, hopefully as far away as possible, could poison the judo ecosystem if online betting and diplomatic stakes were to multiply… In terms of everyday judo, many clubs and even federations would welcome an algorithmic helping hand during the June and September reregistration periods, either to structure payment methods, or to allocate time slots according to age group, for example, not to mention the animation of virtual communities of judokas to exchange current information… Without going so far as to completely replace the tutelary figure of the teacher, AI could also have its say for populations with specific needs or who are temporarily far from the dojo. When asked ahead of the Paris Paralympic Games about ways of preventing cheating in parajudo, orthoptist Vivien Vasseur, coordinator of the clinical investigation centre at the Rothschild Foundation Hospital in Paris, confirmed the future usefulness of AI for validating or invalidating an eye test.

Prefer the I for Intelligence to the A for Artificial. But beware of being blinded by technology. During the mixed team event at the 2024 Olympic Games, the intervention of digital roulette (which, in the event of a three-way tie, determines a weight category called upon to fight a second time to decide between the two teams) caused teeth to cringe on at least two occasions. Firstly, in the very tight round of 16 match between Spain and Japan, the roulette wheel decided which of the possible categories was more favourable to Japan – or more unfavourable to Spain, as the case may be – and won. It was the same story in the final, when the same Japan, this time pitted against the French hosts in a Hitchcock final, saw the category of French legend Teddy Riner chosen to bring the proceedings to a victorious close. In both cases, many commentators, particularly from the defeated nations, cried foul. An unfortunate and slippery slope, but one that should not be taken lightly, lest it spread. If an old-fashioned manual protocol can remove this doubt for good, judo would be wrong to do without it.

The other downside is the human factor. A few years ago, when I was alerted to the repeated non-selections of a French international and the resulting sense of injustice he felt in view of his otherwise good results, I discovered, with a little digging, that the player concerned was not a model of assiduity or punctuality at training – something he was careful not to emphasize – and that keeping him in the squad therefore posed a problem of coherence and overall cohesion for the staff at the time. An important fact that the AI is not (yet) configured to integrate. The article remained in draft form, and that was for the best.

In the beginning. In a short vignette broadcast in autumn 2023 on the Franco-German channel Arte to mark the release of his essay 21 Lessons for the 21st Century, Israeli historian Yuval Noah Harari delineates what he sees as the three great contemporary problems – nuclear war, climate change and technological disruption – and worries in passing that “ethical questions are being transferred from philosophers to engineers“.

Faced with the fantasy of the augmented judoka, half-Vitruvian man, half-Hal 3000 in Stanley Kubrick’s 2001: A Space Odyssey, an article written with a sledgehammer in February 2024, crammed with figures and exclamation marks, by the extremely sharp Thibault Prévost for the demanding French website Arrêt sur images, is responsible for bringing us back down to earth by reminding us of the exorbitant ecological cost of this revolution, which isn’t one for everyone. The cloud is a ‘carbonivorous’ superstructure,” says anthropologist Steven Gonzalez Montserrate, “based on an absurd process: to cool itself, it burns carbon (40% of the electricity consumed by a data centre is burned). This carbon releases CO₂ into the atmosphere, contributing to the global rise in temperature… and making air conditioning ever more essential. […] As the cybersecurity adage goes, ‘there is no cloud, it’s just someone else’s computer’: the magical cloud of data doesn’t exist; there are only the monstrous hangars of Data Center Alley, their kilometres of fibre optics, their inhuman hum.”

Should AI be added to the already considerable carbon footprint of a frequently overheated international circuit which, with the exception of the 2020 containment and the eruption of Iceland’s Eyjafjöll volcano in 2010, has rarely been forced to ease up on its collective consumption of fossil fuels? Greek economist Yanis Varoufakis, recently interviewed about his thoughts on the European Union after the storms his country went through when he was its Finance Minister, smiles when he quotes what Mahatma Gandhi once said about the contribution of British civilization in India: “It might have been a good idea.“

There will undoubtedly be a period of excitement and mutual taming between judo and AI. An aid to decision-making. Support for change. A time to question the meaning and purpose of this tool in relation to Jigoro Kano’s founding intentions. A time to take a step back. To step back. To question. Where does it all start and where does it all go? Remembering that in the beginning was a dojo. A tatami. A teacher. Pupils. And hadjimé. – Anthony Diao.