Did MLB Have the Authority to Ban AI in the Dugout?

MLB did not merely “ban AI” in the dugout. It narrowed what clubs may do on dugout iPads. In a June 11, 2026 memo from executive vice president Morgan Sword, the league restricted those devices, effective July 15, to three non-AI tabs: static statistics, video, and a third function The Athletic described as unspecified. The same report said as many as six clubs had been using AI-generated pitch suggestions from the dugout by midseason, while MLB’s own review found clubs compliant with the rules then in place.[1]

That classification matters more than the device. If the rule is a competitive-integrity measure, MLB starts from a position of real institutional strength. If it is a unilateral change to how players and coaches perform in-game strategic work, the analysis does not end with the commissioner’s authority. It moves into the collective bargaining limitation built into that authority.

The first question is therefore not whether AI pitch suggestions are wise, fair, or aesthetically offensive to baseball tradition. The first question is what legal bucket MLB put the rule in when it acted midseason by operational memo.

The Strongest Case for MLB Starts With Integrity

The commissioner’s “best interests of baseball” authority is not ceremonial. It grew out of the post-Black Sox reconstruction of league governance, when baseball deliberately placed unusually broad integrity powers in the commissioner’s office. MLB.com’s account of the clause traces that lineage from the 1921 constitutional settlement through later commissioners who used the power in matters tied to confidence in the game, including Kenesaw Mountain Landis, Bowie Kuhn, and Bud Selig.[2]

That history gives MLB its cleanest argument. AI-generated pitch recommendations in the dugout are not just another spreadsheet. They can influence pitch selection, sequencing, matchup choices, and substitution timing while the contest is underway. A league need not wait for a sign-stealing scandal, a device-security breach, or a club-by-club technological arms race before deciding that real-time algorithmic assistance belongs outside the dugout.

The best version of MLB’s position is preventive rather than punitive. If six clubs were already using AI-generated pitch suggestions, the league could reasonably see the competitive environment beginning to tilt toward tools that are difficult to audit and easy to expand.[1] A rule that limits dugout tablets to static statistics and video can be framed as a line-drawing exercise: clubs may prepare, study, and review, but they may not bring live algorithmic prescription into the bench area.

That is not a frivolous integrity theory. League offices routinely regulate equipment, communications, replay access, wearable technology, and bench conduct because competitive rules are meaningful only if they can be administered in real time. The fact that the technology is analytically sophisticated does not make it immune from league control. Innovation, by itself, is not a legal entitlement.

Nor does the absence of proven misconduct eliminate MLB’s authority to act. A commissioner’s integrity power would be an oddly weak instrument if it could operate only after the league had already suffered the confidence problem it was designed to prevent. On this side of the ledger, MLB can say it identified a category of in-game assistance that was approaching a governance boundary and clarified the boundary before the practice became entrenched.

The Carve-Out Is Where the Analysis Gets Harder

The problem is that “best interests” is broad, not bottomless. Article II, Section 2 of the MLB Constitution gives the commissioner sweeping authority, but it also makes that authority inapplicable to collective bargaining matters. That carve-out prevents the analysis from stopping at the traditional recital that commissioners have wide discretion over integrity.

AI-assisted pitch suggestions are close to the line because they affect the work being performed during games. A catcher receiving signs, a pitcher choosing a sequence, a bench coach preparing a matchup move, and a manager deciding whether to change pitchers are not merely using club property. They are carrying out the strategic labor of the game. A rule that removes a tool from that process can be characterized as a rule about how bargaining-unit employees and field staff perform their in-game functions.

MLB will resist that framing. It can argue that the memo regulates club technology, not player employment terms: no player loses pay, roster status, service time, discipline protection, travel conditions, or medical rights because an iPad tab disappears. On that theory, the rule sits with operational restrictions on what information may be available in the dugout, much like limits on communications systems or in-game video access.

The union-side answer is stronger than a complaint about inconvenience. If the league changes what information a catcher may consult, what recommendations a pitching coach may relay, or what analytical support may be used before a substitution, it is changing the decision environment in which players and coaches work. The National Labor Relations Act treats working conditions and mandatory bargaining subjects broadly enough that a technology rule can become a labor rule when it materially affects job performance, evaluation, or discipline exposure.

That distinction is easy to understate because “dugout iPad policy” sounds administrative. But in-game information architecture is not neutral. It shapes who has authority in the moment: the catcher, the pitcher, the bench, the advance analyst, or the model. Removing AI-generated suggestions may restore more discretion to players and coaches, or it may simply replace one set of practices with another. Either way, it touches the conditions under which strategic baseball labor is performed.

The Signature Block Matters

The memo’s posture gives the labor argument more room. The rule reportedly came in a June 11 memo from Morgan Sword, not as a commissioner-level integrity order, and took effect July 15.[1] That does not make it invalid. League executives act for the league every day. But the form of the action says something about how MLB appeared to classify the problem when it acted.

If MLB believed AI pitch recommendations created an immediate threat to the integrity of competition, one might expect a more formal invocation of the commissioner’s integrity authority. The league instead used the machinery of operational administration. That choice is useful for implementation, but it is less useful if the dispute becomes one over unilateral change. A routine technology memo is easier to characterize as a management-rights adjustment than as an emergency integrity intervention.

The compliance finding also matters. According to The Athletic, MLB found clubs compliant with existing rules.[1] That fact does not deprive the league of power to regulate prospectively. It does, however, weaken any emotional borrowing from scandal-response precedents. Black Sox, Finley, Steinbrenner, and Biogenesis sit in a tradition of commissioner action tied to confidence in the game, but a prospective ban after a finding of existing compliance occupies a quieter category.

That quieter category may still be lawful. It is just not the same category. The more MLB presents the rule as ordinary administration of dugout devices, the more natural it becomes to ask whether the administration affected a mandatory bargaining subject. The more MLB presents it as integrity regulation, the more it must explain why the record supported immediate unilateral action when clubs were reportedly following the rules then in force.

Timing Pulls the Rule Toward the Bargaining Table

The calendar is not a technicality. The current collective bargaining agreement runs through December 1, 2026, and the AI restriction took effect in July, with the next agreement already in view. A midseason change made during an active bargaining cycle is not automatically unlawful, but it will be read against that procedural background.

That timing creates two different narratives. MLB’s narrative is that AI use was emerging quickly, with several clubs already deploying pitch-suggestion tools, and that a leaguewide rule was needed before inconsistent club practices hardened into competitive imbalance.[1] The MLBPA’s likely narrative is that the league saw a bargaining issue approaching and chose to settle it unilaterally before the subject could be traded, narrowed, or conditioned in negotiations.

The latter framing does not require the union to prove that AI is good for players. A bargaining claim would not turn on whether pitchers prefer model-generated suggestions, whether catchers resent them, or whether coaches believe the tools improve decisions. The question would be whether the rule changes terms or conditions of employment enough that the league had to bargain before imposing it.

That is why the narrowness of the ban cuts both ways. MLB can point to the remaining tabs and say clubs still have static statistics and video, so the rule merely removes a discrete category of automated recommendation.[1] The union can answer that the removed category is precisely the one that changes the in-game decision process. A static chart is reference material. A model-generated pitch suggestion is advice directed at the next competitive act.

The Minor-League Pattern Is Relevant, But Not Controlling

MLB’s December 2025 move to centralize minor-league technology regulation belongs in the background, not the center of the major-league analysis. It suggests a broader institutional preference for bringing baseball technology under league control before club-by-club variation becomes difficult to police. That pattern is unsurprising. Once technology becomes embedded in player development, game planning, and real-time decision support, retroactive governance is messy.

But major-league dugouts sit inside a different labor architecture. A policy that may be administratively clean in the minors does not answer whether a major-league rule affects collectively bargained work conditions. The more direct legal materials remain the MLB Constitution’s best-interests grant, its collective bargaining limitation, the CBA calendar, and the actual function of AI pitch suggestions during games.

Where the Authority Likely Lands

MLB probably has a defensible authority argument. A league charged with protecting competitive integrity can regulate real-time algorithmic assistance in the dugout, especially when the practice is already present among multiple clubs and the rule is framed as a uniform limit rather than club-specific discipline.[1] The commissioner’s historical mandate is broad enough to support prospective governance of technology that affects competition.[2]

The legal exposure is not that MLB acted in an area wholly outside league concern. It is that MLB acted unilaterally in an area that can plausibly be described as strategic work performed during games. Once the rule is understood that way, the constitutional carve-out for collective bargaining matters becomes more than a footnote. It becomes the boundary between integrity regulation and a bargaining obligation.

That boundary will not be resolved by slogans about human judgment or technological progress. The relevant distinction is institutional fit. If the rule is treated as a competition rule governing what information may enter the dugout, MLB’s footing is relatively strong. If it is treated as a unilateral change to the tools players and coaches use to perform in-game strategic labor, MLB carries more risk into the December 1 CBA deadline.

References

1. MLB bans AI-generated pitch suggestions from dugout iPads, The Athletic, July 16, 2026.
2. Richard Justice: Best interests of baseball a wide-ranging power of commissioner, MLB.com.