Context
The Government is applying AI to improve efficiency and effectiveness across Whitehall. A concern raised by the Post Office scandal involves human civil servants making mistakes that can be explained, whereas AI systems may make significant errors without providing an explanation.
Question
When a human civil servant—let us say at His Majesty’s Revenue and Customs or the DWP—makes a mistake and is challenged, they can explain their logic and how they came to the decision. We know that the courts always believe that computers are best and give the right answer, but AI makes mistakes—sometimes huge ones. Because of the way it is programmed, it cannot explain how it got to the decision. How will the Government ensure that the appeal process continues to work and we do not have a high-tech version of the Post Office scandal?
Answer from The Chancellor of the Duchy of Lancaster (Pat McFadden)
The right hon. Gentleman raises an important issue. The public inquiry into the Horizon scandal shows that blind faith in a computer system used in a court of law can lead to injustices. I do believe in the possibilities of AI, but it is important to keep the human element at all times. It will enhance human productivity but not replace it. That is the way we should go.
Not addressed: The Chancellor did not provide specific measures or guarantees for maintaining an appeal process when using AI in public services.
How: Discussed Principles Without Concrete Actions
Was this summary accurate? Tell us
How accurate was this summary? (1 = poor, 5 = spot on)