AINEWS
The league told teams to stop using generative AI for in game strategy calls, a rare mid season directive that draws a line on who decides in the dugout rather than neutralizing a real competitive edge.
MLB told teams this month to stop using AI chatbots on dugout iPads for substitutions, pitch calling, and other in-game calls. The ban does not claim the technology was actually moving the needle, and on the evidence available, it wasn't.
A June 11 memo from the commissioner's office, first reported by The Athletic on July 16, said iPad AI use had "gone beyond their originally intended purpose to include recommendations regarding substitutions, pitch calling, and other in-game decisions traditionally made by players and coaches." Anonymous front-office sources told The Athletic that roughly a third of MLB clubs were querying AI chatbots on the tablets for in-game strategy input. The league reviewed all 30 clubs after the memo and found them compliant; no team is being punished. Enforcement began after the All-Star break on July 22.
Mid-season directives are rare in MLB. The most recent comparable move was the 2021 mid-season crackdown on foreign-substance glove checks on the pitcher's mound. Treating an iPad-AI policy with the same urgency signals how seriously the league is drawing this line.
The available evidence suggests the AI wasn't actually winning many of those calls. A June 2026 study referenced in Futurism's coverage found large language models are weak at predictive sports analytics tasks, even when they handle perception-style jobs like identifying players on court. Industry commentary from WSC Sports makes the same point: off-the-shelf LLMs are not high-accuracy forecasters unless wired into live data streams. The chatbots teams were running on dugout iPads almost certainly weren't that.
MLB is not regulating a tool that was demonstrably shifting outcomes. It is drawing a normative line about who decides in the dugout, before the technology gets good enough that the line would be much harder to hold. The ban reads less like a competitive correction and more like a preemptive preservation of the human-decision chain.
Moneywise reported that the Mets had invested in a "very expensive" custom AI program, relaying unnamed sources, which would put them among the clubs most affected by the new rules. The Athletic's anonymous sourcing puts the broader adoption figure at roughly a third of the league.
MLB has shown it is willing to integrate technology into officiating when the boundary is clear. The league adopted the Automated Ball-Strike challenge system for 2026, letting hitters and pitchers request a robot umpire on individual pitches. The ABS system is bounded: it is built into the rule book, it is transparent, and the human players still decide when to use it. Dugout AI chatbots, by contrast, sit outside any rule frame and inside the coaching workflow. The new memo is the league's first attempt to write that frame.
Teams have already shown they will push the technology into whatever edge they can find. The next pressure points are bullpen phones, dugout-to-rotation communications, and any generative tool the league has not yet named. The ban took effect on July 22.