Professors use Balogun red card to frame AI limits in decision-making
Columbia Business School professors argued that soccer’s use of replay and tracking tools shows where AI can decide and where humans still must judge.
By Daniel Okafor · Business Editor
3 min read
Columbia Business School professors Oded Netzer, Christopher Frank and Paul Magnone said Folarin Balogun’s disputed red card at the World Cup offers a business lesson about artificial intelligence and human judgment. Writing in Fortune, they argued that technology can settle some decisions cleanly, while leaving the hardest calls more exposed and more dependent on people.
The professors pointed to two first-half incidents from the United States men’s national team’s July 1 Round of 32 match in Santa Clara. In one, semi-automated offside technology ruled out what appeared to be a Balogun goal after showing he had moved ahead of the last defender by a narrow margin, they wrote.
That call drew little dispute, according to the professors, because the question was measurable. Cameras, sensors and AI-assisted systems could determine where the player was when the ball was played.
The second incident was different. About 30 minutes later, Balogun’s trailing boot made contact with Bosnia defender Tarik Muharemović’s leg and ankle, the professors wrote. The video assistant referee flagged the play, and referee Raphael Claus reviewed slow-motion footage at the sideline monitor before showing Balogun a red card for serious foul play.
The United States went on to win with 10 players, but the sending-off became widely debated, according to the professors. Their point was not to decide whether Claus made the right call, they said, but to show that better technology does not remove the need for human expertise in contested decisions.
Routine calls versus judgment calls
Netzer, Frank and Magnone separated decisions into two broad categories. Some are binary and data-driven, such as offside decisions or goal-line rulings in soccer, where technology can answer the question with little need for interpretation.
They compared those calls with business tasks that can be automated, including sending an incoming request to the right queue, choosing an advertisement for a user or flagging a transaction that matches fraud indicators. In such cases, they argued, human review can slow the process and add error rather than insight.
The professors cited a survey of C-suite executives that found 44% would change a planned decision based on AI insights. They said the value of that AI input depends on the type of decision involved: a measurable call can be handed to systems more readily than a decision involving context, risk, proportionality or intent.
More data can sharpen scrutiny
The professors argued that more evidence often narrows the set of easy calls while making the remaining decisions more visible and contested. In soccer, they wrote, replays can show contact from several angles, but they cannot settle whether a challenge was reckless, accidental or deserving of dismissal.
For corporate leaders, they said the task is to decide which decisions are resolved once reliable data arrives, and which ones still require accountable human judgment. They warned against both over-reliance on automated systems and resistance to AI output when the data is well suited to the question.
Netzer, Frank and Magnone, who co-authored Decisions Over Decimals: Striking the Balance between Intuition and Information, said the spread of AI changes the role of decision-makers rather than eliminating it. Their conclusion was that referees and executives face a narrower set of decisions, but the calls left to them demand greater judgment.
This story draws on original reporting from Fortune.