When New York Mets owner Steve Cohen reportedly spent several hundred thousand dollars on an artificial intelligence program that helped coaches pick pitches and make in-game strategy calls, the Mets did not gain an edge for long. Major League Baseball moved swiftly to shut down the practice league-wide, and by the start of the second half of the 2026 season, every iPad in every dugout across the league had its custom application tab locked. The policy change is one of the first major professional sports league regulations targeting real-time AI use by teams, and it sets a precedent that founders in any competitive, regulated industry cannot ignore.
The move came through a June 11 memo from MLB executive vice president of baseball operations Morgan Sword, who wrote that the custom tab on dugout iPads had been used beyond its original purpose. The memo, obtained by the Associated Press, stated that teams were running AI programs delivering recommendations on substitutions, pitch calling, and other in-game decisions traditionally made by players and coaches. MLB made the tabs inaccessible starting July 15, when the season resumed after the All-Star break.
How a Several Hundred Thousand Dollar AI Program Triggered a League-Wide Policy Change
The chain of events began with Adam Ottavino, a former Mets reliever who now broadcasts for the New York Yankees on the YES Network. On his YouTube livestream Baseball and Coffee, Ottavino revealed that the Mets had been the primary target of MLB's investigation. The team had deployed an AI program that Ottavino described as expensive, and Mets personnel had been discussing it openly within the league. Some of the coaches I know were talking about it from around the league, and they had basically an AI program helping them pick pitches, Ottavino said on the stream.
The revelation sent shockwaves through baseball. Yankees captain Aaron Judge told reporters he could not believe teams were making decisions off AI. Toronto Blue Jays manager John Schneider called it a little weird to see how decisions could be swayed in real time by technology. The message from the league office was unambiguous: the boundary between human judgment and algorithmic recommendation had been crossed, and the line needed to be redrawn.
MLB's investigation, conducted through its competition committee, found that clubs had been compliant with the new restrictions after the deadline. But the fact that it took a specific complaint and a formal memo to shut down a practice that had clearly been running for months reveals a broader truth about AI adoption in competitive environments. By the time a regulator catches up, the technology has already been deployed.
The Timeline From iPads to AI Crackdown
MLB's relationship with dugout technology has evolved in stages. The league began a pilot program allowing iPads in dugouts with restrictions late in the 2015 season and expanded their use in 2016 under a deal with Apple. After the Houston Astros sign-stealing scandal, where the team used video cameras and replay monitors to decode opposing catchers signs, video was eliminated from dugouts during the pandemic-shortened 2020 season. It returned in 2021, but the Astros scandal had already made the league acutely sensitive to how technology could undermine competitive integrity.
The difference this time is the technology itself. The Astros used a live camera feed and a human relaying signals. The Mets AI program was different: it processed game data, analyzed pitcher tendencies, and output pitch recommendations in real time without waiting for a human to make the call. That distinction matters. A camera feed requires a human intermediary. An AI recommendation engine skips the intermediary and goes straight to decision support. The line between tool and decision-maker had blurred, and MLB decided to draw a bright line against it.
Arizona Diamondbacks manager Torey Lovullo captured the tension perfectly in his response to the policy change: AI is entering everyones arena. You better get on it, or you are going to get rolled over by it. It is a strikingly honest admission from a sitting MLB manager that the genie is not going back in the bottle. The league can restrict iPad tabs today, but the technology itself is not going away.
What This Means for Founders Building in Regulated Environments
MLB's AI crackdown is not just a sports story. It is a case study in how regulators across every industry will respond when they discover AI tools operating in spaces they did not anticipate. The pattern is already repeating: the EU AI Act phases in throughout 2026 and 2027, the White House issued a security directive forcing Anthropic to disable Fable 5, and now a professional sports league is restricting AI use in dugouts. Each of these interventions follows the same arc. A technology emerges, a few early adopters push it further than expected, and the governing body responds with restrictions that reshape the competitive landscape.
For founders, three lessons stand out. First, the timeline between deployment and regulation is shrinking. The Mets AI program appears to have been running for months before MLB acted. But the gap between initial use and formal restriction will only narrow as leagues and regulators become more vigilant. Second, early adopters absorb the scrutiny. The Mets did not receive a penalty in this case, but they were outed in public, their competitive advantage revoked, and the policy change was written around their behavior. Third, the regulated path becomes a moat for compliant players. Once MLB defines what is allowed, the teams that had invested in compliant solutions will have a head start on whatever the league permits next.
The most revealing line in this entire story came from Sword's June 11 memo. The custom tab had expanded the use of the dugout iPads beyond their originally intended purpose, he wrote. That language is going to sound familiar to any founder who has ever received a written warning from a platform, a regulator, or a marketplace. The phrase beyond intended purpose is how every regulatory action begins. The question is not whether your industry will face similar scrutiny. It is whether you will be the Mets or the league office when it arrives.
What Founders Need to Do
If you are building AI tools for any competitive, regulated, or professional environment, here is the checklist to run through before your industry writes its version of MLB's memo. Audit what data your AI consumes during real-time decision-making. If your model ingests live game, trading, bidding, or operational data and outputs recommendations within seconds, you are operating in the same territory as the Mets AI program. Identify which governing body has authority over your product category. For MLB, it was the competition committee. For your industry, it could be the SEC, FINRA, the FDA, or a self-regulatory organization. Have you mapped who writes the rules? Document every recommendation your system makes. When a regulator asks what your AI suggested and when, you need a log that shows the exact inputs, the timestamp, and the output. Without that, you have no defense. Build a compliance mode into your product from day one. The teams that shrugged off the iPad restriction were the ones that had never used the custom tab. The teams that scrambled were the ones that had built their workflow around it. Design your AI so that a regulator can flip a switch and bring you into compliance without breaking your product. And finally, assume that what is allowed today will be restricted tomorrow. The most durable AI companies will be the ones that treat regulation not as a surprise, but as a feature of the landscape they were already planning for.
