AI is already changing the Olympics. Starting with the Milan Cortina Games

One year ago, in April 2024, the International Olympic Committee (IOC) presented its Olympic Agenda on Artificial Intelligence in London: a strategic document destined to shape the future of sport and the entire Olympic Movement. It was not just a declaration of intent: behind it were months of intense work with experts from universities such as Oxford and MIT, together with global technology partners such as Intel, Alibaba and Deloitte.

Objective? To integrate AI into the Olympic Games in a responsible way , ensuring fairness, transparency and sustainability. That Agenda has already left its mark. At Paris 2024, AI monitored over 2.3 million social conversations to protect the mental health of athletes, contributed to more accurate judgements in competitions and enabled more sustainable management of venues thanks to digital twins . But the Committee's gaze is already set on Milano Cortina 2026 , where the challenges are even more complex: alpine environments, high-speed winter disciplines, growing needs for digital safety and competitive fairness.

To understand how AI is transforming sports, both on and off the field, we interviewed Ilario Corna , Chief Technology and Information Officer of the IOC. We talked with him about AI-assisted judges, new tools to protect athletes online, energy sustainability and technological governance, in a journey that aims not only at more modern Games but also at leaving a lasting digital legacy for the entire Olympic Movement.

The Olympic Agenda on Artificial Intelligence, presented in April 2024, indicates that AI must be implemented “in compliance with Olympic values, ensuring fairness, transparency and accountability”. In concrete terms, what priorities are you setting for Milano Cortina 2026 to combine support for athletes’ performances, protection of their mental health and safeguarding the integrity of the competitions?

“At the Olympic Winter Games Milano Cortina 2026 we are developing AI solutions that put athletes at the center, in full alignment with Olympic values. One of our key priorities is equitable access: we want AI to be a tool and an opportunity for all International Federations and National Olympic Committees, not just those with more resources. For this reason, we will continue to offer free training and share AI tools globally. On mental health, we will expand the AI-based social media monitoring system, already used during the Paris 2024 Games, which analyzed over 2.3 million posts and flagged more than 10,200 offensive comments for removal. This had a direct impact on the protection of athletes. To ensure the integrity of the competition, we are working on AI-supported tools for objective, transparent and accountable judging. All developed in close collaboration with the International Federations.

In Paris, in collaboration with Omega, computer vision tools have refined the accuracy of real-time judgments. What specific advances are you developing for winter sports – for example, alpine skiing or short-track skiing – and what technical challenges remain for AI to become a true “second judge” at Milano Cortina 2026?

“We are testing innovative solutions, such as helmet-mounted cameras for short track skaters. The footage, uploaded to cloud platforms, will be analyzed by AI to help judges determine whether falls were caused by irregular contacts or accidental loss of balance. The goal is to support judges by analyzing critical racing incidents - such as skate blade proximity, physical contacts and the causes of accidents - with AI-enhanced videos. Key challenges include reducing latency, athlete safety and seamless system integration. We are working with the International Skating Federation to finalize prototypes that meet rigorous safety and performance standards.”

You announced that new refereeing and data analytics solutions were tested at the ISU World Tour in Milan last February. What key indicators (accuracy, latency, athlete and federation acceptance, etc.) are you using to evaluate their success and decide on a possible permanent adoption at the Games?

“Within these pilots, including the Milan stop of the ISU World Tour, we are monitoring several key performance indicators: data accuracy, real-time latency, usability for judges, acceptance by athletes, officials and federations. Transparency is also key: all data must be verifiable and easily interpretable. While mounting a camera on an athlete’s helmet may seem simple, safety remains a top priority. In collaboration with the ISU and a motorsport engineering company, we are applying rigorous safety methodologies to ensure the technology poses no risk in competition. Our approach starts with tools in ‘0.1’ or ‘1.0’ versions, which we refine through an iterative process based on real-world feedback. Only when these technologies meet rigorous performance and safety criteria do we consider them for Olympic adoption. A major benefit of partnerships with international federations is the ability to test these innovations at official events prior to the Games. This collaborative model is essential to accelerate progress while maintaining trust and transparency.”

At Paris 2024, AI monitoring of 2.3 million posts led to the automatic removal of over 10,000 offensive content, as mentioned. How do you intend to scale this system for the social ecosystem of the Winter Games - where the volume of conversations and languages ​​could be different - and what ethical guarantees have you foreseen to avoid false positives or unjustified censorship?

“The system will be adapted to the specific context of the Olympic Winter Games: different languages, smaller volumes and variable sports visibility. But the main objective remains the same: to protect athletes and staff from online abuse.

We have implemented strong ethical safeguards to ensure responsible use of AI: automated systems do not make removal decisions autonomously, especially in sensitive cases. Human moderators are always involved to assess critical content. One lesson from the Paris Games is that language models sometimes miss contextual nuances that are obvious to humans. For example, when asked “What are the rules for these Games?”, AI did not always understand that “these Games” referred to the Paris 2024 Games. To address this, we are improving context understanding, especially around event-specific terminology and time references.”

The “Digital Twin” experience in Paris has shown how AI can reduce costs and environmental impact of venues. How are you combining energy simulations, predictive weather models and transport planning to minimise emissions from Milano Cortina 2026 while ensuring mobility for athletes, spectators and media in complex Alpine contexts?

“We are significantly expanding the use of Digital Twin technology, which has demonstrated great operational value at Paris 2024. This technology enables the creation of dynamic, data-driven simulations of facilities, transportation and environmental conditions such as temperature, wind, precipitation well before the start of the Games. These virtual models help organizers identify logistical issues in advance, optimize operations and reduce unnecessary travel, contributing to more sustainable planning and lower environmental impact. In parallel, we are increasing the use of the Alibaba Cloud Energy Expert platform, already used in Paris to identify energy waste such as lights left on in empty areas or HVAC failures and provide real-time corrections. At Milano Cortina, we will integrate predictive weather forecasts specific to Alpine environments, in order to anticipate peaks in energy demand based on conditions such as heavy snowfall or sudden drops in temperature. This will enable timely decisions on heating, snow removal logistics and facility management, improving energy efficiency and operational resilience in a context where the climate is variable and often unpredictable”.

The IOC is already working with partners like Alibaba, while Milano Cortina is in talks with TIM for 5G networks and edge cloud. How are you harmonizing data security standards, algorithm intellectual property, and internal training—mentioned in your AI courses for staff and federations—to ensure these technologies remain a lasting legacy for the Olympic Movement beyond 2026?

“Our strategy is based on three pillars: security and privacy standards, intellectual property protection and development of skills within the Olympic ecosystem. We are collaborating with global technology leaders such as Alibaba and, for Milano Cortina 2026, with partners such as TIM to advance 5G connectivity. All projects are guided by the principles of our Trustworthy AI Framework, which ensures ethical, legal and operational safeguards are built in from the start. For example, when developing solutions based on LLM (large language models), we work closely with legal and compliance teams to ensure we are aligned with emerging regulations such as the European Union AI Act — including requirements on data governance, transparency, traceability and proactive risk management. A key focus is education: we offer free AI training to all National Olympic Committees and International Federations to build lasting internal capabilities. This ensures that skills remain in the Olympic Movement, fostering autonomy and innovation beyond the Games. Finally, we remain committed to transparency: all our tools, frameworks and AI results are made public. The goal is not only to optimize Milano Cortina 2026 but to provide the Olympic Movement with the tools to adopt AI responsibly and consciously. AI is here to stay. The opportunity we face is not only to adapt but to lead change, leveraging this transformative technology to strengthen the integrity, inclusiveness and long-term impact of the Olympic Games and the global community they represent.”