As an entrepreneur and tech enthusiast, I have witnessed several overhyped technologies and businesses. These stretch from the first wave of the internet in the 1990s with Webvan and Pets.com, which both had multibillion-dollar valuations, to the recent Theranos scandal, where a $10bn blood testing business turned out to be a sham.
Irrational exuberance has been the precursor to the downfall of many ventures. I have been far from immune; you only have to see the photo of me, proudly wearing my Google Glasses, sitting next to the ponytailed inventor Astro Teller in 2013 as evidence. However, I believe that OpenAI’s ChatGPT, an artificial intelligence tool, could be a gamechanger. Bill Gates recently declared it the most significant technological advance since the graphical user interface.
Since ChatGPT’s public release last year I have been exploring its usefulness at home and in businesses. I encourage my 11- and 14-year-old children to use ChatGPT as a personalised learning assistant. In the business world I suggest using the tool in meetings, with a smart employee framing questions to help the collective meeting arrive at better conclusions. Using these tools may provide a competitive advantage through early adoption, at least in the short term.
In 2015 I attended a talk at IBM about the capabilities of Watson, its flagship AI. One of the developers I spoke to said something that stuck with me: that a better way to think about AI would be for us to think of it as IA or “intelligence augmented” – a set of tools and capabilities that will not replace us but enhance our own human capabilities. It has yet to have a visible impact on sport.
In football ChatGPT can easily be used to create marketing and communication content but that is like using a Formula One car to transport your shopping. The more interesting question is how AI could enhance insights and create advantages in the game. Many top clubs have data scientists and analysts working on player performance to gain an edge, from recruitment to training, diet and match analysis. The use of AI in football will likely be a closely guarded secret but it is undoubtedly part of the success stories of Tony Bloom and Matthew Benham, the visionary owners of Brighton and Brentford respectively.
Rumours suggest that they have armies of “quants” looking to find undervalued players in markets worldwide, similar to Moneyball. Machine learning (ML) will already be part of their business interests and will be based on the players continuing their trajectories over the following decade and establishing themselves on the European football scene.
Football, like most sports, is a combination of art and science. For most of its history it has been an art but now there are clear advantages to incorporating science. The first wave of technology and insight has led to huge improvements in physical and tactical capabilities in the home camps of those using progressive tools and insights. ChatGPT is shortening the distance between the data and its utility, making it accessible and more useful. Before a match teams use a video analyst to evaluate the strengths and weaknesses of the opposition. However, this system has inherent flaws and is dependent on the analyst’s experience and insight. AI could recommend the optimal training sessions before the game to exploit the tactical and physical vulnerabilities of the opposing team.
In future, all historical data of live games could be used to recommend ways to line up and play against the opposing team. During the game it could be possible to layer that data in real time to receive recommendations on how to adjust strategy. For example, computer vision could identify that an opposition right-back generally tires and loses pace in the 73rd minute by observing capillary dilation in their face; the AI would recommend adding a fresh left attacking winger. Analysis could identify that a goalkeeper might tend to drop high balls more frequently in the first nine minutes of a game and recommend sending high balls in the first 10% of play. It could be that high crowd noise, above 90 decibels, might cause certain teams to lose concentration so let’s find a way to increase the crowd noise. The hypotheses are endless.
This may sound far-fetched, but in an article in Wired magazine Liverpool recently announced a collaboration with Deepmind to “combine computer vision, statistical learning and game theory to help teams spot patterns in the data they collect”. It is already possible to get historical game data but under a rumoured new media deal that would allow for live streaming of all games, the visual data would be available for all clubs to access immediately. The tooling to analyse that data would not be difficult, because HD cameras and computer vision software could track individual players.
Generative AI and ChatGPT could be the last piece of the puzzle to create broader adoption and widen the use cases. In 10 years head coaches could have an AI assistant advising them on formations and substitutions. Although the game will always represent the complex, multi-variant interactions of 11 individual players there may be insights and advantages that AI can provide about those interactions. Liverpool may not be reaping the benefits this season but this early technological move could be important to their long-term success.
We are early in the life of these technologies and should not be misled by their perceived linear development. In the coming years the combination of these technologies with new hardware such as quantum computing could set us off on an exponential curve affecting not just sport but every area of our lives. In 2018, I had the opportunity to sit next to Professor Yoav Shoham, a world expert in AI, at a dinner in Tel Aviv. I wondered how close we are to achieving truly intelligent machines capable of full‑scale “general intelligence” similar to or better than humans.
Shoham shared a story about a cartoon he saw when he was young, which depicted a child standing on a little chair while looking at the stars through a telescope. He used this image to describe the current state of AI, where the stars represent general intelligence and the chair symbolises current AI machine learning. With the advent of large language models used to train ChatGPT, one wonders if we have now climbed up on to the table but could soon be on our way to the stratosphere.
Jason Stockwood is the chair of Grimsby Town
