At the 2026 Winter Olympics in Milan and Cortina d’Ampezzo (Milano Cortina 2026), Olympic Broadcasting Services broadcast AI-generated replays and real-time performance analysis. According to an estimate by Yiannis Exarchos, chief executive of Olympic Broadcasting Services, more than half of the Olympic audience consists of casual spectators, and this is the audience that technology is primarily targeting. The same tools, in different form, are widening the gap between federations that can finance data science and those that cannot. Milano Cortina 2026, and even more so Los Angeles 2028, raise the question of who benefits from these advances, and on what terms.
The Essentials
- Sports technology pursues two simultaneous objectives that benefit distinct groups: increasing the performance of elite athletes and enriching the experience of casual spectators.
- At the 2026 Olympics, Olympic Broadcasting Services deployed AI replays and real-time analysis; more than 50% of Olympic television viewers are not sports fans (source: Axios, February 2026).
- In leagues like the NBA, access to analysis platforms creates a measurable competitive advantage that reinforces existing hierarchies rather than flattening them.
- Philosopher Jean-Michel Besnier identifies in this movement a deeper temptation: the athlete’s body is treated as an optimizable system, where finitude and unpredictability become bugs to fix, not conditions of merit.
- LA 2028’s AI integration will be even more ambitious, making it urgent to determine whether this infrastructure can be shared, or whether it is structurally reserved for those who finance it.
A Two-Tiered Sport That Technology Does Not Create, But Accelerates
A divide has long existed between nations and clubs that possess the budgets, infrastructures, and medical staffs necessary, and those that do not. Technology did not invent it. It measured it with new precision, then deepened it.
The mechanism is well-known in innovation economics: tools that increase productivity tend first to benefit those who can adopt them quickly, widening the gap before closing it, sometimes. The digital labor market offers a direct illustration, as data on AI-related training analyze: access to skills largely determines who benefits from the wave, and who suffers from it.
In sport, the cycle follows the same logic, but with an aggravating factor: results are measured in fractions of a second, in centimeters, in world ranking points that determine subsequent funding. A national federation that gains access to real-time biomechanical analysis before another does not merely have additional comfort. It possesses a cumulative advantage: better injury prevention, finer tactical adjustment, more targeted recruitment. The advantage capitalizes.
The NBA illustrates this clearly. Since the widespread deployment of optical tracking systems in arenas, Second Spectrum, then proprietary platforms of franchises, teams equipped with robust analytics departments have systematically overrepresented the top of standings. The Houston Rockets, then the Golden State Warriors, then the Boston Celtics each, in their time, transformed an analytical lead into athletic domination.
Television Broadcasting Masks This Asymmetry
Watching the Milano Cortina 2026 Olympics from your living room is to see technology in its most visible and accessible version: an AI replay that reconstructs a skeleton fall in 360 degrees, a split-screen mode with AI camera following national athletes in biathlon with real-time shooting data, a graph comparing the trajectories of two skiers at the entrance to a turn. Olympic Broadcasting Services designed these tools for the general public, and they work. With more than half of Olympic television viewers being casual spectators, according to the estimate cited by Axios, this technological layer serves as an initiation interface: it translates performance into something readable for those unfamiliar with skeleton rules.
But this public visibility masks another reality. The same analysis tools, deployed on the athlete and staff side, are not equally accessible to all delegations. A national federation with a contract with an embedded data science provider receives real-time data streams during competitions, exploitable for immediate training decisions. A federation without this partnership watches the same television replay as the average viewer, with at best a video analyst running the file the next day.
Technology thus creates two audiences and two levels of use for the same event. This duplication is structurally inscribed in the economic model. Olympic Broadcasting Services sells broadcasting rights and develops products for the general public. Analytical solutions intended for teams are marketed separately by private actors, according to pricing logics that mechanically exclude nations with fewer sports resources.
The Athlete as System, or Finitude as an Undesirable Variable
Jean-Michel Besnier, a philosopher specializing in technology and digital humanism, questions in his latest work the extent to which we want to resemble the intelligent objects we construct. His thesis, developed in N’être plus qu’un objet, is that contemporary technophilia reverses classical alienation. It is no longer merely that machines imitate humans: we want to behave like them, be hyperconnected, measurable, optimizable, repairable. Consciousness, finitude, unpredictability become handicaps to correct.
Applied to elite sport, this framework is strikingly relevant. The Olympic athlete augmented by data science is precisely treated as a system: their physiological parameters are continuously captured, their biomechanical weaknesses identified algorithmically, their recovery cycles modeled. Performance becomes the result of optimization, not transcendence. The fall, the technical error, the collapse under pressure—everything that constitutes the properly human texture of competition—becomes input data for a predictive model rather than a manifestation of the condition of the subject who runs or jumps.
Sports medicine and biomechanics have always sought to correct and optimize, and video analysis has existed since the 1970s. The rupture is quantitative rather than qualitative: it stems from a saturation of measurements rather than an absolute discontinuity. When every millisecond of a movement is decomposed, when every parameter is modeled in real time and when an algorithmic recommendation can modify an athlete’s technique between two rounds, the share of autonomous decision, embodied judgment, and body intelligence shrinks accordingly.
The response varies depending on what one expects from sport. For a pragmatic coach or a performance economist, the tool is legitimate insofar as it improves results, and merit remains with the athlete. For Besnier, the response is more cautious: the tool transforms the relationship the subject maintains with their own performance, and by extension with themselves.
Real Possibilities Opened by Data
The pessimistic reading is too convenient. The history of technology in sport is not solely that of the elite consolidating its lead. It is also that of the progressive diffusion of tools that reconfigure practices at all levels.
The case of cycling is instructive. Power meters, reserved for professional teams in the 1990s, are now worn by tens of millions of amateurs worldwide at accessible prices. The biomechanical knowledge resulting from this has improved practice at all levels, not just at the top. Tennis experienced the same movement with ball tracking tools and video analysis platforms, now available to amateur clubs in Europe and North America.
Technology diffuses structurally, even if the delay can be long. The number of Olympic cycles that elapse between adoption by wealthy nations and availability to others remains the central variable, and active mechanisms could accelerate this diffusion.
On this point, LA 2028 represents a test. The AI project for the Los Angeles Games is described by Olympic Broadcasting Services as significantly more ambitious than Milano Cortina 2026, with deeper integration of performance data into mass broadcasting. If the International Olympic Committee chooses to centralize and share certain analytical infrastructures, as it already does for television broadcasting, it is plausible that medium-sized delegations would gain access to data they could not have afforded individually.
This scenario is not guaranteed. Commercial logic pushes in the opposite direction: private technology partners have an interest in selling differentiated solutions to the most solvent delegations. But there is a precedent in other fields where inequalities of access to digital tools have been partially mitigated by active sharing policies, a question also arising in industry, where access to capital and innovation follows similar logics of initial concentration.
The Spectator Between Initiation and Distraction
The other side of the divide deserves to be examined for itself, without cynicism. Reaching more than 50% of non-fan television viewers is a significant sporting and cultural fact. These people would probably not watch the Games without the layer of explanation and staging that technology enables. If AI replays and data visualizations lead them to understand what they are watching, to distinguish the technical quality of a figure in figure skating from what superficially resembles the same thing, something is transmitted.
One can debate the nature of what is transmitted. Steven Pinker and researchers in progress science would insist on the value of a broader sports culture, of increased participation even at a distance, of making accessible performances once reserved for those who could afford a ticket to Milan or Cortina d’Ampezzo. The criticism Besnier would formulate would be different: if the spectator sees sport only filtered through analysis algorithms, if performance is delivered to them already digested, glossed over, translated into metrics, they consume a narrative about the body rather than a presence of the body. The emotion is real, but its object is mediated in a new way.
This tension does not necessarily call for a verdict. It signals that making access to sporting spectacle and understanding of sporting performance more accessible are two distinct things, that technology can serve one without necessarily deepening the other.
Unresolved Stakes for 2028
Los Angeles 2028 is less than two years away. Technological architecture decisions are being made now: which data streams will be shared, which tools will be available to which delegations, how will innovations in mass broadcasting be articulated with training tools.
The precedent of Milano Cortina 2026 shows that AI integration in service of the spectator can work at large scale, and that the non-fan public finds something in it. Milano Cortina 2026 did not resolve the question of access equity on the athlete side, and available sources do not allow it to be settled. If LA 2028 deploys an even more ambitious analytical infrastructure without explicit sharing mechanisms, the gap between technologically equipped delegations and others will be larger than it has ever been.
The Olympic movement must decide whether it treats analytical infrastructure as a common good of sport, in the same way as the rules of the game or the anti-doping protocol. Letting the market decide exposes it to a result where 2028 medals reflect data science budgets of delegations more than the value of the athletes wearing the bibs.
Sources
- Axios, “AI is playing a bigger role at the 2026 Olympics,” February 18, 2026: https://www.axios.com/2026/02/18/artificial-intelligence-2026-olympics-replays
- Jean-Michel Besnier, N’être plus qu’un objet : La tentation d’oublier la vie, Éditions Hermann: https://www.editions-hermann.fr/livre/n-etre-plus-qu-un-objet-jean-michel-besnier
- Olympic Broadcasting Services, AI deployment data Milano Cortina 2026 (cited via Axios, link above)
- Syracuse University Sport Analytics Lab, analysis of the impact of analytical data on NBA competitiveness (cited without direct URL, source mentioned in the brief)
