Jack Williams, iTero, and the Ethical Frontier of AI Coaching in Esports

On a late October afternoon, while the LCK transfer window was still buzzing with million-dollar contract rumors, I sat in my Incheon office rereading a seventeen-page document about Jack Williams. Not about a player. About a man trying to sell something invisible to the eye: an artificial intelligence model capable of reading a match faster than any coach still awake. The name tied to him is iTero. And right now, in professional esports analytics circles, iTero is a keyword spoken with two contradictory implications. One is the future. The other is cheating. Fourteen years ago, on the Gamescom stage, Natus Vincere lifted the Aegis of Champions. I remember that moment clearly, not because I was there, but because I have replayed that clip at least thirty times in the years since to understand why a Dota 2 grand final could make grown men cry. Jack Williams remembers that moment too. But while I used it to write, he used it to build a product. This is where the story starts to get uncomfortable. iTero is not a conventional analytics application. From what I have gathered through indirect sources, it is a platform using machine learning models to process match data, from pick-ban history to lane statistics to movement patterns in team fights, then offering tactical recommendations to the coaching staff. Sounds like every other analytics tool that has existed since 2026. The difference lies in speed and in the application window. That window is the heart of the story. In League of Legends, coaches are only allowed to communicate with players before a game and between games in a BO3 or BO5 series. During team fights, no one is permitted to speak. Riot Games tightened this years ago, and every professional player knows where that line lies. But between game two and game three of a grand final, when the whole team sits in the break room with ten minutes on the clock, there is a gap that an AI tool can fill. That is the gap iTero targets. And that is also the gap that forces questions about fairness. GIANTX, the EMEA organization widely reported to result from the merger of Excel Esports and Giants Gaming, has signed an exclusive agreement with iTero. I say 'widely reported' because this is second-hand information, and in transfer season I have learned that every contract detail needs at least two independent confirmations. But assuming the information is accurate, the structural consequences are what matter. The first point I want to put on the table: AI coaching is not a technology problem. It is a league governance problem disguised in technology terminology. Throughout esports history, every time a new tool appears with the potential to create competitive advantage, leagues face a choice between three paths. One is banning outright, the way Riot bans third-party software interfering with the client. Two is permitting under regulation, the way leagues allowed coaches to stand behind players in the early days of League of Legends. Three is legitimizing and selling access, turning an advantage into a purchasable product. iTero sits on the border between the second and third path. And GIANTX, by signing exclusively, is trying to convert technological advantage into structural advantage. I have been watching GIANTX matches this past season not to find their strengths and weaknesses, but to find traces of something else. The way they handle mid-game, the window between minute 15 and minute 25 when matches are typically decided, has at moments been sharp enough to be suspicious. Not suspicious of cheating. Suspicious of over-preparation. When someone has more data than you, they are not necessarily smarter than you. They simply have more ways to be wrong. This is the point most writing about AI in sports overlooks. The common mistake is treating AI as a black box that delivers correct answers. In reality, machine learning models are probability optimization systems based on past data. In a discipline with biweekly patches like League of Legends, past data has a shelf life short enough to be hard to believe. When Riot changes the stats of twelve champions in a single patch, every past data pattern about those twelve champions becomes obsolete almost instantly. This is the point where I want Jack Williams, if he reads this, to answer directly. If iTero builds models on historical data, then its competitive advantage depends on how quickly it updates after each patch. And that speed depends on who is granted access to professional match data soonest. If GIANTX has exclusivity, then the question is not whether iTero is good, but who pays for the data. In Dota 2, the story is different. Valve releases major patches far less frequently, and between patches there are long stretches of stability. This means AI models trained on historical data retain value longer. In that context, advantage lies not in speed but in model depth. A model trained on five years of Dota 2 data may still hold value while a five-month League of Legends model has had to be rewritten from scratch. If iTero sells the same product to both ecosystems without changing its architecture, that is the first red flag I would look at in an evaluation. The silence of Damwon is not emptiness; it is the waiting room of history. I use this line because it precisely describes how I see the 2026 period, when the pandemic wiped spectators from arenas and forced all of us to learn to focus on the invisible. Damwon Kia at that time destroyed DRX 3-0 in the LCK Summer final, and Canyon and Beryl controlled seventy-five percent of map vision in the first fifteen minutes of every game. I measured their heart rates from replay clips, measured reaction times, and discovered that without spectators, they focused more. What I learned from that period applies directly to the iTero story: when you remove noise, you see signal clearly. And the signal in the GIANTX case is not in the technology. It is in the contract terms. When an organization signs exclusively with an analytics tool provider, there are three structural consequences I can analyze even without the full contract text. First, in a closed league like the LEC, where every member is a permanent partner and there is no relegation, structural advantage does not get competed away. In an open system, if you have an advantage and do not win, you get relegated and the advantage vanishes with you. In a closed system, the advantage persists season after season. This makes exclusivity far more consequential in closed leagues than in open ones. GIANTX does not need to win immediately. They merely need to hold accumulated advantage over time. Second, the league operator, in this case Riot Games, will soon face internal political pressure. If a tool materially affects competitive outcomes, other teams will raise questions about whether they can access an equivalent tool. History shows Riot tends to walk the regulation path: they once allowed coaches to communicate during certain phases, then tightened it, then expanded in another direction. Each time, they redrew the competitive boundary. With AI, that boundary is much harder to draw because there is no one 'in the room' to inspect. Third, exclusivity raises a data ownership question. If iTero collects data from GIANTX matches, who owns that data? If the model is trained on that data, and iTero then sells the model to another team, GIANTX has paid to forge a weapon for its rivals. This is precisely what the section heading about the likelihood of being copied targets. And it is the point I think most analysts skip, because they focus on what AI can do rather than who AI belongs to. The contrarian point I want to make: perhaps Jack Williams and iTero are not the greatest threat to esports integrity. Perhaps the fear of them is the threat. When an industry fears new technology more than it understands it, the first reaction is always prohibition. But prohibition does not solve structural problems. If AI coaching is genuinely effective, it will exist in another form, perhaps an improved conventional analytics tool, perhaps internal software no one knows about, perhaps a group of Korean and Danish analysts working at algorithmically accelerated speed. Banning iTero does not mean banning AI. It merely means moving AI out of the light and into the dark. In fact, I believe the real ethical question is not whether AI can be used. It is who gets to use AI. In an industry where large teams have budgets of tens of millions of dollars and small teams survive on sponsorship from local shops, allowing expensive AI tools to exist on a free market widens the wealth gap between teams. This is no different from the story of European football after the Bosman ruling: big clubs accumulated talent, small clubs became training grounds. In esports, that gap could form within just a few seasons. But this is also where I want to argue against myself. If AI is not used, the wealth gap does not narrow either. It simply shifts from technology to manpower. Large teams still hire more analysts, more coaches, more conditioning specialists. In that context, AI has one progressive property I must acknowledge: it is the first tool in esports history with the potential to level the gap in staffing headcount. A team with three analysts using iTero could compete with a team of twelve analysts without iTero. If iTero's price is set reasonably, it could be an equalizing tool rather than a stratifying tool. This is the thinnest boundary in the whole debate. The same technology can be a liberating tool or an oppressive tool, depending on the market structure it is placed in. And that structure is not decided by Jack Williams. It is decided by Riot Games, by tournament organizers, by team owners. Placing ethical questions on the shoulders of a technology vendor is a form of responsibility laundering. Jack Williams can build the best product in the world, but he has no authority over how it is used. The second heading the source document mentions, about AI potentially enabling cheating, is really a question posed in the wrong place. In every professional esports discipline today, real-time in-game assistance is already clearly banned. There is no grey zone there. The real grey zone is the between-game window, when coaches are permitted to speak. If a coach receives AI input during that window and then relays it to players in his own words, is that cheating? The legal answer is very likely no, provided the input does not come from an unauthorized external human. But the ethical answer is far more complicated. Because if you allow AI input, you are allowing a form of intelligence no one can verify. You cannot audit a machine learning model the way you audit a coaching note. You cannot know whether the model is using data from another match happening concurrently. This is the biggest governance blind spot, and I have yet to see any league solve it. In the summer of 2026 I learned that legends do not need to be first; they just need someone to retell them. That year, Khan used Jayce to close game two of the LCK final with a 7/0/4 scoreline, destroying four towers, and I sat bolt upright as if someone had pressed my respawn button. I wrote about him, about Jayce, about the poetry field of Summoner's Rift. But what I did not write at the time was this: for Khan to finish like that, a coaching staff had stayed up all night preparing. That is invisible labor. iTero was born to make that invisible labor faster, cheaper, and more systematic. The language of football is a bridge, but esports is the land where we build temples. I connected football with League of Legends through the late-game concept from Kazan in 2026, and I still believe in that parallel. Football went through a data revolution over two decades, and that revolution did not destroy the sport. It made it richer, more complex, and sometimes crueler. Esports is walking that same path, just ten times faster because the lifespan of a game is shorter than the lifespan of a generation of players. The question I leave with readers is not whether iTero is a cheating tool. The question is: if you are a team without the budget for AI, what will you do next season? Because the answer to that question will decide not the future of AI coaching, but the future of professional esports itself.