The future of sports content: How AI transforms the live experience

Sports, at their essence, are physical endeavours: competitions between athletes that dramatically unfold on a pitch of finely mowed grass, or a freshly waxed court.

This will never change. What we see unfold before our eyes, from youth sports to the Olympics, from college sports in North America to the World Cup, inspire us to see what literal heights humans can reach whilst leaving us aspiring to be involved in those competitions like never before. We may not actually be able to play at a high level, but we want to be engaged with as much of the detail as possible.

Over time, we have strived to better understand every aspect of the games we love, by documenting what we saw in person. First, as a scoreline for the newspaper, then through radio and the beginnings of data capture with statistics, then a video broadcast followed by more detailed eventing and tracking. Across all levels of sport, there is a desire to translate the physical match into a comprehensive digital record of every game.

Today, data is at the heart of sports. It is key to delivering insights to coaches and players that impact on-field performance, it helps officials monitor and control the game, and it fuels fans’ ability to engage with and enjoy the action.

While effectively collecting data is a critical first step, the ability to transform data into the detailed language of the sport, insights, and products is what unlocks real value. Unfortunately, in many competitions today, this value is not realised because data is siloed, not synchronised, too slow, and not delivered to the market in a form that can be easily utilised.

At Genius Sports, we are using AI to create the ultimate digital replica of a physical match. At every moment, sports and their partners can now access a live parallel digitised match that mirrors what is being played out on the physical pitch. Every flick of the foot, jab step, curl of the ball can be instantly replicated in the digitised world. A fan can experience all of this in any number of ways: stats, stories, broadcast TV, augmented streams, or AR/VR experiences. But it doesn’t end there. Once digitised, all of this information can then be fed back and used to help execute the physical game.

Possession of the ball can be instantly decided when the ball goes out of bounds. An official can be instantly alerted when there is an offside call, or a tackle that is a potential foul. Head coaches can be alerted to game trends on the bench and adjust strategies in real-time. The list goes on.

Creating a machine understanding of sport

To bring this vision into reality, we have developed state of the art AI capabilities in three main areas, resulting in a machine understanding of sport:

1. Dragon tracking: State of the art computer vision tracking systems: Our computer vision-based Dragon tracking system, capturing 10,000 surface mesh data points per player, over 200 times per second, or billions of data points in an individual game.

2. Insights: Machine learning applied to tracking data automates event collection: Our machine learning systems translate the billions of Dragon data points into the language of sport. In real-time we contextualise every aspect of the game, using the language of coaches, players, and fans.

3. Personalised fan experience: Enabling immersive broadcasts: Our computer vision team processes broadcast video and synchronises it with our data collection. This combination allows us to augment live broadcast video with a limitless set of data and graphics, enabling truly personalised fan experiences.

Crucially, and core to our vision at Genius Sports, this technology is not to be the preserve of elite sports alone. Every level of sport should – and eventually will – benefit from the AI revolution. This is why we designed Dragon, and all the downstream use cases it powers, to scale up to the highest levels of competition, but also scale down to local recreational play. This is the future of sports. Everywhere.

AI is transforming fan experiences

This technology is transforming how fans experience sports and how leagues and rights holders deliver content.

Many of you reading this will have seen some form of live augmented broadcast, with graphics and data being used to tell stories within the game or activate sponsorships in eye-catching ways.

These augmented experiences are only possible by synchronising tracking data with real-time video. AI needs to both understand the context of what is happening in the game and know exactly where and when to surface a graphic or a data point within a video feed. And to do this live, the whole process needs to be automated as there is no time for a producer to manually insert a visual.

Augmentation can power alternate or themed broadcasts or be used to support new immersive storytelling within the main presentation of a game. Early use cases have included stat-heavy ‘altcasts’ by Prime Video and brand activations by the likes of Nickelodeon and Marvel.

Premier League Productions, through their Data Zone product, is also innovating in this space. Through our partnership with the Premier League and their global broadcasters, Data Zone serves up a host of real-time insights, comparisons and data visualisations throughout the game, from a dynamic pitch map to shot speeds above the goal, runs in-behind, progressive passes, integrated fantasy data and much more.

AI is transforming professional workflows

When machines can understand sport, at the level of, or beyond, the capabilities of a professional coach, it fundamentally changes the workflows for the most serious practitioners of the game.

Rather than spending countless hours manually collating data about a game, our AI automatically indexes every event in real time, and precisely links it to video. Team staff can shift their time from data collection and alignment, to working with and coaching their athletes.

At the touch of a button, analysts can create a game-by-game breakdown on any facet of team or individual performance such as a Premier League winger’s 1v1 dribbles which can be further broken down to when the player attacks the defender’s left or right side. Analysts and coaches can ask any question about the game, and have answers delivered to them in data, visualisations, and video playlists with just a few keystrokes.

An AI-powered future

Alternate viewing modes and breakthroughs in team and player performance analysis are just beginning. There is no limit to how many ways we can present and enhance a single live sporting event.

In time, each fan can ultimately experience a game in a way that is perfectly tailored to them.  An AI that understands the game and understands a fans preference, can deliver what a fan wants, when they want it, automatically.