← AI for Educators Daily with Dan Fitzpatrick
Loading summary
Transcript1 lines
- [00:00]
A
If this episode makes you think, please let us know in the comments and support us by subscribing and leaving a review. Thank you. Today we are exploring AI governance lessons from a football containing a sensor that sampled its movement 500 times every second. The article is what a Chip in a Football Teaches us about governing AI, published by the World Economic Forum and written by Jake Okeechukwu Efadoo, assistant professor at the Lincoln Alexander School of Law at Toronto Metropolitan University. Effadue uses one fiercely disputed decision from the 2026 World cup to make a wider argument. Better detection doesn't automatically create fairer decisions. Toronto, 2 July the 103rd minute Croatia's Josko Gvardiol forces the ball over the line against Portugal and apparently equalizing in a knockout match. Referee Espenescos is sent to the pitch side screen. But the decisive evidence isn't a conventional replay. It's a waveform from a sensor inside the ball showing an extremely slight contact with Igor Mitanovic, perhaps involving only his hair. That touch placed another player offside. The goal was disallowed, Croatia were eliminated and FIFA said the contact had been proven. The technology may well have been accurate, yet many supporters experience the decision as unfair. That distinction matters. A sensor can establish whether contact occurred. It can't tell us whether a barely perceptible touch should carry enough significance to decide a World cup match. That's a question about the rules and the values behind them. I found this fascinating because discussions about AI governance often concentrate on whether a system produces an accurate result. Efidu's argument is that accuracy is only one layer. We also need to ask whether the rule being applied is reasonable, whether similar cases receive similar treatment, whether people can challenge a decision, and whether the institution operating the system has earned public trust. Those are useful AI governance lessons for schools. Suppose an AI system identifies students considered at risk of missing their expected grades. It could be statistically accurate, but what happens next? Who defined at risk? Is the prediction based on attendance, behaviour, previous assessment, family circumstances or some combination of those? Can a teacher challenge it? Can the student see the reasoning? Does the label create support or quietly lower expectations? The model might measure exactly what it was designed to measure, while the surrounding decision remains deeply flawed. The article is also careful about language. Not everything described as VAR in football is artificial intelligence. The semi automated offside system used cameras and computer vision, meaning software that identifies and tracks visual information. It compared player positions with three dimensional scans of all 1248 players. The Adidas Trionder ball contained an inertial sensor measuring movement, video assistant, referee or var, remained a process in which humans reviewed footage under rules set by the International Football association board. So AI in sports is really a collection of sensors, automation, data analysis, interfaces and human judgment calling all of it. AI can hide where authority actually sits. We do something similar in education, especially school may say An AI system recommended an intervention when the system merely generated a score. A person chose the data, another person selected the threshold, and a leader decided what consequence followed. Responsibility can dissolve behind the phrase, the system says. This is where the familiar idea of keeping a human in the loop needs more scrutiny. I strongly support human oversight, but Effidue is right that the phrase is incomplete. Which human is in the loop? What do they know? What authority do they have? Are they genuinely reviewing the output or simply approving it? Because the machine appears precise, a polished dashboard can create undeserved confidence. In football, frozen frames, slow motion and three dimensional avatars can make an interpretive decision look like a scientific fact. In a school, a traffic light display can do the same. Red feels objective, green feels safe, yet the colors usually reflect human choices about categories and thresholds. If you value this kind of practical analysis of AI in education, follow or subscribe so you don't miss future episodes. The source gives a particularly uncomfortable example. England's Jarell Quansah reportedly received a two match ban following a VAR assisted red card with without a route of appeal during the tournament, the article says. Officials were criticized for initially showing the referee a frozen image rather than the full movement. By contrast, FIFA suspended the United states player Folerin Balogan's automatic ban after President Donald Trump reportedly lobbied FIFA's president, according to the article. No published reasoning accompanied that reversal, and UEFA said the intervention had crossed a red line. The same technology existed in both cases. The surrounding access to influence and remedy did not. That's the strongest part of the piece for me. Technology enters institutions that already contain uneven power. It doesn't arrive on neutral ground. Automating a process can make existing rules faster and more consistent, but it can also accelerate their defects. As Effadue puts it, football upgraded its eyes without upgrading its constitution. For a school leader, that means procurement can't end with questions about accuracy, privacy and price. Those matter, of course, but leaders also need to map the whole decision system. What happens when the tool is wrong? Who can override it? Is that override recorded? Can a parent or student receive an understandable explanation? Are remedies consistent? Or do confident and well connected families gain more influence? Think about a department head planning professional development around AI assisted marking. The tempting session is about using the tool efficiently. A better session might give teachers the same student response and the same AI feedback. Then ask them to identify where professional judgment changes the outcome that makes disagreement visible and useful. It teaches staff to separate measurement from judgment rather than treating the software's wording as a verdict. In a Year eight geography lesson, students could examine the Yoshko Gvardiol decision and design an appeals process. They'd have to decide what evidence counts, who gets to review it, how quickly a ruling must happen, and whether technically correct decisions can still be disproportionate. That's richer AI literacy than simply learning to prompt a chatbot. Students are learning how automated evidence interacts with rules, institutions, and power. The policy implications become more pressing because the European Union Artificial Intelligence act, commonly called the EU AI act, brings major obligations for certain high risk systems into force. From August 2026, the article focuses on hiring, welfare, credit, and border security. The conversation around the AI act and education needs the same discipline Compliance paperwork alone won't produce legitimacy. Governments can mandate human oversight, but unless staff have time, training, authority, and a meaningful review process, oversight becomes ceremonial. I don't think the article argues against the technology, and neither do I. Semi automated offside systems removed some obvious errors, public announcements and referee body cameras Improved transparency. Spain may have been crowned World cup champions amid controversy, but greater precision still has value. The mistake is expecting precision to carry a burden it cannot carry. Machines can help detect People must still interpret, justify, and remain accountable. Human in the loop should mean more than a person clicking approve. What I keep coming back to is this. Before automating a school decision, review the rule. Separate evidence from judgment. Create a genuine route of challenge and and make responsibility visible. Measure consistency and impact alongside technical accuracy. Those are AI governance lessons worth carrying from VAR in football into every education system. Considering automated decisions, a precise system can enforce a poor decision perfectly. So govern the whole decision, not merely the machine. That's all for today. Thanks for listening.