AI Is Rewriting Cricket — Without Replacing the Humans Yet

Cricket’s AI revolution has already begun

Robots capable of bowling a perfect doosra remain a distant prospect, but artificial intelligence is already changing how cricket is analysed, how young players are developed and how the sport is covered.

Freddie Wilde, an analyst working with teams in the Indian Premier League and the Hundred and head of cricket intelligence at CricViz, says “almost all elements” of his daily work have been touched by AI. The technology is not yet replacing players, coaches or analysts. Instead, it is allowing them to work more independently, handling tasks that once required data engineers or data scientists.

For batters, that can mean identifying where a bowler is likely to pitch and deciding which risks are worth taking. For bowlers, it can reveal weaknesses against particular angles, lengths or field settings. The final decision, and the execution, remain human.

From performance data to fairer pathways

One of the most significant applications is emerging away from elite match analysis. EquiPlay, a talent-pathway platform, standardises coaching assessments, tracks player development and uses AI to turn scores into narrative feedback.

The platform has recorded more than 2,000 assessments involving 300 young cricketers, completed by 28 coaches in a first-class county pathway. Prof Adeela Ahmed Shafi of the University of Gloucestershire and a board member at Gloucester CC says the objective is not to replace coaching judgment but to make it visible.

That visibility could help clubs identify whether players from different ethnic, socioeconomic, educational or geographical backgrounds are progressing at different rates. If state-school players begin to fall behind at a particular age, for example, clubs may be able to design an intervention. The system can also test whether coaches assess players consistently or favour particular groups.

“Before, we wouldn’t have noticed that,” Shafi says. “We knew who was coming into the pathway, but we never really knew how they were progressing.”

AI cannot easily measure culture, character or the balance of a dressing room, however. Those judgments remain human — at least for now.

The line between assistance and substitution

The debate becomes more difficult when AI moves from supporting decisions to making them. Cricket has accepted technology-assisted umpiring through systems such as the Decision Review System because a human remains at the centre of the process. A system that made every decision perfectly could challenge that arrangement.

The central question is not simply whether AI can outperform people, but whether the sport would accept its authority. At what point does assistance become substitution?

Developments in robotics offer a warning about how quickly expectations can change. At China’s second World Humanoid Robot Games, more than 500 robots competed in sports including football, athletics and boxing. Many stumbled or needed human help, but one reportedly completed 100 metres in 8.64 seconds.

What happens to cricket journalism?

AI is also reshaping the media. It can produce a report in seconds, but that speed creates a new responsibility: journalists and editors must interrogate, refine and challenge the material rather than simply publish it.

The danger is stylistic as well as factual. If prose, jokes and observations are drawn from the same pool of language, cricket writing could become more efficient but less distinctive, with fewer individual voices and more interchangeable accounts of the sport.

Live cricket retains an important defence. As Wilde puts it, live sport is “authentic and real”, unpredictable and impossible to stage. Data and planning matter, but execution, uncertainty, memory and meaning cannot be reduced so easily to a model.

AI may help cricket understand itself more clearly. It may make pathways fairer and analysts more powerful. But the sport’s enduring appeal still rests on human choices and imperfect moments — precisely the qualities that no machine-generated summary can fully capture.