Artificial intelligence: authoritarian by default?
The global tennis circus has decamped from London for another year, leaving behind not only the usual dramatic memories but also an artificial intelligence (AI) controversy with relevance far beyond Wimbledon.
After 147 years, those traditionalists at the All England Lawn Tennis and Croquet Club (AELTC) dispensed with line judges at their Championships this year, replacing them with a system they call Electronic Line Calling (ELC).
ELC works by using 12 concealed cameras around the court to track the ball’s movement, which a computer then interprets to determine whether it lands in or out, triggering a call announced by a recorded human voice.
The new system seemed to have won overwhelming support from players and spectators — until it failed during a match because of ‘operator error’. At a key moment there was no call of ‘out’ when a ball landed rather obviously beyond the base line, leading to an intervention by the umpire. Having consulted someone by phone, he ordered the point to be replayed.
“It’s really important to say that the umpire followed the protocols in place,” AELTC CEO Sally Bolton explained later. “He did what he needed to do on court and acted entirely correctly.”
Yet, in the past, an umpire noticing such a clear error by a human line judge would have awarded the point to its rightful winner rather than ordering a replay, and that’s where this incident becomes of wider interest.
ELC is surely an example of a job better done by tech than by a human, but the change in umpire authority is an example of how over-reliance can disempower human judgement.
This mindset matters well beyond the trivial example of tennis. If so much respect is vested in AI that a previously empowered umpire is no longer authorised to trust the evidence of his own eyes, what will happen when AI is bolted on to organisational systems that already centralise power and undervalue the judgement of the people closest to the action?
None of what happened at Wimbledon, or in any other situation in which AI is replacing human judgement for better or worse, is the fault of the tech itself, although that is easily overlooked when the discourse treats it as a superhuman power.
This tendency to deify AI – or, indeed, to satanize it – is itself an avoidance of the truth that its impact on us will be determined not by the tech itself but by power relationships between us humans.
So long as power and resources are concentrated with the few rather than dispersed more democratically – whether at governmental or organisational levels – we should expect AI to be authoritarian by default even if not by design.