スポーツの動画レビュー論争が示す AI と人間の意思決定

If you watched the 2026 FIFA World Cup, you probably saw the now-familiar ritual: a goal, a celebration, then a referee touching an earpiece while fans and players wait for the verdict from the video assistant referee, or VAR.

That ritual is not unique to soccer. Baseball has a replay center in New York, and other sports have also moved once-human calls into review systems.

The hope is simple: Use better technology to yield better calls.

Soccer authorities expected more cameras, replay angles and tracking data to mitigate referee error, limit subjectivity and strengthen perceptions of fairness. But VAR has created new arguments among players, coaches, fans and commentators: Should VAR have intervened? Was the same standard for when to intervene applied consistently? Did the process itself feel fair?

Watching the World Cup as a decision-making researcher, I could not help but see VAR as more than a refereeing tool. It looked like a case study in technology-assisted decision-making, with parallels to the growing use of AI in organizational decision-making.

VAR produced a set of mismatches between what the technology promised and how people experienced it. It left fans disappointed and created new disputes about judgment, process and trust. These tensions offer a useful lens for understanding AI adoption more broadly.

Better measurement is not better judgment

Some decisions are measurement problems: Was the player offside? Did the ball cross the line? Did contact occur? Technology is excellent at these questions because cameras, sensors and data reduce mistakes that come from people not seeing clearly.

But many decisions are judgment problems: Was the contact enough for a penalty? Was the tackle reckless? Was the referee’s original call clearly wrong? Those questions involve interpretation, context and standards. A camera can show that contact occurred. It cannot decide how that contact should be judged.

The same issue appears when organizations use AI to support decisions. AI can process data, predict patterns and classify cases. But a doctor, manager, judge or teacher still needs to decide what the output means, how much weight to give it and what values are at stake. Research on human-AI collaboration makes a similar point: AI is often strongest when paired with human judgment rather than treated as a full replacement for it.

Subjectivity does not disappear – it moves

Before VAR, arguments focused on what the referee saw. After VAR, arguments often focus on how the process works. When does VAR intervene? How far back can officials review the play? What counts as “clear and obvious” video evidence?

Subjectivity is still there. It has moved from the referee’s eyes to the rules, thresholds and governance of the review system. Baseball shows a similar pattern: Replay did not remove judgment from the game. It changed which plays could be reviewed, how challenges worked and where judgment entered the process.

AI systems create the same shift. Organizations often imagine AI as a way to remove human discretion. In practice, discretion reappears in new places: deciding which model to use, what data to train on, what error rate is acceptable and who is accountable when the system fails.

More precision can reduce trust

People often assume that greater precision creates greater trust. Sometimes it does. But it can raise expectations faster than it reduces ambiguity. If technology can measure an offside decision by centimeters, fans may expect every decision to feel equally certain. When penalty calls or red cards remain debatable, people can become frustrated.

Research on why people avoid using algorithms finds something similar: People can lose confidence in algorithms after seeing them make mistakes, even when they perform well overall. Accuracy alone does not guarantee legitimacy.

That finding matters for AI decision systems. A company may use AI to make hiring or performance evaluation more objective. But if applicants or employees see the system as inconsistent, biased or unfair, trust can fall instead of rise. When organizations frame AI as a silver bullet, they raise expectations the technology cannot always meet. The result can be deeper frustration and faster loss of trust.

Where technology ends and judgment begins

Management research suggests that AI adoption today is not a simple choice between humans and machines. The better question is whether a decision should be automated, augmented or left to human judgment.

Measurement problems are the strongest candidates for automation: AI can scan documents, detect patterns, classify cases or flag anomalies faster and more consistently than people can. Interpretation problems call for human-AI collaboration: AI can provide information, options or recommendations, but people still need to think about the context of the decision, how much uncertainty is involved and the consequences the decision may have. And some decisions remain deeply human. Questions involving fairness, accountability, values or meaning cannot be handed over to technology without changing the nature of the decision itself – from one of human judgment to algorithmic assessment.

VAR makes this boundary visible in sport. AI is now forcing organizations to confront the same boundary across high-stakes decisions. The key question in AI adoption is not simply, “Can AI make this decision?” It is, “Which parts of the decision should AI make, and which parts should remain a matter of human judgment?”