Geoffrey Vos: AI could judge injury, negligence cases

Vos's comments come as AI is already changing behavior in the courts. He points to a 10 percent rise in county court cases started every year, which he links to people using tools like Gemini, ChatGPT, or Claude to draft claim forms. While some of those claims may be bogus, Vos argues it is a mistake to assume they are all frivolous, and that the increase shows greater access to justice.

The debate over AI adjudication also runs into questions of trust. Vos notes that people may find it harder to accept an adverse decision from a machine than from a human judge, and that any move toward algorithmic decision-making would require consent in a particular case or a societal decision through legislation. On deepfakes, he is less alarmed than some, describing fabricated images and recordings as an extension of false evidence courts have always had to weigh, with perjury laws still applying.

With Vos retiring in a few days, the question of where to draw the line is left to others. He frames it as a matter for society, politicians, and legislation rather than judges alone.