Full disclosure: I’m a life-long Wolverhampton Wanderers fan, and all my fellows of the Old Gold will attest that we’ve had some tough VAR rulings against us – particularly this season. And having lost our chance at playing in the Wembley tie on Sunday to Coventry City, I was quietly backing them versus Manchester United.
We all know the score after penalties, and you’ll have struggled to miss the debate surrounding the 121st minute ‘City goal being disallowed by VAR. And with all the controversy surrounding VAR, and working in the AI field myself, on my Monday morning commute I couldn’t escape musing “Should we replace VAR with AI?”
Somewhat tongue in cheek, perhaps, but here’s what I’d suggest we’d need to think more about…
- Accuracy and Consistency: The primary reason for considering AI over human-operated VAR is the potential for increased accuracy and consistency in decision-making. AI can be programmed to make decisions based on predefined rules and criteria without human biases. It works in industry.
- Technical Feasibility: Implementing AI for VAR requires advanced technology capable of real-time analysis of game footage. This includes computer vision and machine learning algorithms to interpret game events accurately. The technology is there. And we see it in other sports already.
- Complexity of Decision-Making: Football refereeing often involves subjective judgments that are not solely based on video evidence. Decisions can be influenced by context, intent, and the dynamics of the game. AI systems would need to be sophisticated enough to understand these nuances. Again, they already are.
- Ethical and Regulatory Considerations: Using AI for critical decisions in a high-stakes environment like football raises ethical questions. Who is responsible for the decisions made by AI? What if there are errors or biases in the AI system? But the human “trial” is seeming rather unbalanced, so what’s the real risk?
- Public Perception and Acceptance: Fans, players, and managers might have mixed reactions to replacing human-operated VAR with AI. There could be concerns about the ‘human element’ being lost from decision-making. This is the one I’m probably most drawn by, but again – the “professionals” seem to be making a right muddle of things so far.
- Testing and Development: Before implementing AI in live matches, extensive testing and validation are necessary to ensure reliability and fairness. Training AI systems with vast amounts of historical match data would be crucial. But with the past few seasons’ data readily available and failures captured, we could have AI learned and running for the start of the next competition or season – time is not against us.
- Hybrid Approaches: Instead of replacing VAR entirely with AI, a hybrid model combining AI technology with human oversight might be more feasible. This approach could leverage the strengths of both AI and human referees. Again, this is happening in other sports already. It does leave us still open to contention, but perhaps less so?
- Adaptation Period: Introducing AI in such a critical role would require an adaptation period for players, coaches, officials, and fans to become accustomed to the new system. Maybe better to “trial” with a “smaller” competition and gain critical mass of support, but with a growing voice to kill VAR entirely, perhaps this ship has already sailed?
I know I’m opening myself up here: there will be readers and comments from all sides – some likely in conflict. But I do think it makes for a fascinating debate. Do we let the machines help solve a man-made problem, or take the bravery pill and opt for a roll-back to how it was in the good old days, before VAR?
Internet, it’s over to you….
About DCS
We are a Pure-Play Pega Partner and team of decisioning experts providing advisory, implementation and on-going managed services for enterprise clients globally. Working in partnership with enterprises, we help leverage the power of Pega’s AI, next-generation AI and decisioning capabilities to deliver 1:1 Customer Engagement, Credit Risk and Intelligent Automation solutions. With offices in the UK, Netherlands and India coupled with planned expansion into the US and APAC markets our team of over 100+ Decisioning Consultants have expertise across a range of industries – including Financial Services, Communications & Media, Insurance and Retail sectors.
