MLB draws a line on AI iPads in dugouts

MLB has banned teams from using iPads in dugouts for apps that make live in-game recommendations, including substitutions, pitch selection, and other decisions typically handled by coaches and players.

The move, first reported by The Athletic, was communicated to clubs in a memo from the commissioner’s office on June 11. According to the report, up to a third of teams had already been using tablets not just as reference tools, but as platforms for custom software offering real-time recommendations during games. MLB did not punish any clubs after reviewing the practice, saying teams had adjusted before the rule took effect.

The wording is narrow. The ban applies to applications that provide “recommendations on substitutions, pitch calls, and other game decisions.” MLB is not banning data analysis outright. Baseball has relied on models for years in scouting, opponent preparation, and pitch-by-pitch evaluation. The league’s new position is that pregame analysis is one thing; an in-dugout screen acting like a second manager is another.

Dugout tablets themselves are not new in MLB. The league began allowing digital tools in place of paper printouts and bulky stat binders more than a decade ago, giving teams quick access to video and hitter-pitcher splits. Apple has long been a prominent technology partner in that setup.

Recommended reading

Hassabis says STEM makes you 10x better at AI

Signal-stealing fallout and the new AI boundary

MLB has become much more restrictive about digital tools near the field since the sign-stealing scandals, especially the Houston Astros case, which led to a $5 million fine and executive suspensions in 2020. The logic is straightforward: the more technology sits close to live play, the greater the temptation to use it for more than analysis.

This latest AI restriction follows that same pattern. The league still embraces technology where it controls the process. Two clear examples are:

• Statcast, the tracking system used to measure ball speed, launch angle, and defensive performance
• ABS, the automated balls-and-strikes system MLB has tested in the minor leagues and at exhibition venues

So the issue is not algorithms themselves. It is where MLB believes measurement and officiating end, and coaching begins.

That debate extends beyond baseball. Teams in the NBA and NFL also rely heavily on advanced analytics, but the public line has usually been the same: models can inform, while the final in-game call stays with the coaching staff. Generative AI blurs that boundary by producing ready-made answers in conversational form, immediately, on the bench.

There is also a practical side. Baseball still depends on factors that do not always fit neatly into a model: a pitcher’s feel, a catcher’s nerves, a reliever’s fatigue. With the average MLB franchise now worth more than $2.6 billion, according to Forbes, any disputed technology practice quickly becomes both a competitive issue and a reputational one. The next real test will come if a team challenges the line between analysis and recommendation once the new season starts.