Context
The question is prompted by the Government's recent initiatives to improve technology use in public services, including a written statement on using AI to boost productivity.
Question
I welcome the written statement from the Minister for Digital Government and Data on the plan to use artificial intelligence to boost productivity in public services. I am concerned about gender and racial bias in artificial intelligence, which many studies have shown, particularly as regards health outcomes. A King’s College London study showed racial bias in AI when it comes to heart scans, indicating worse health outcomes for black and minority ethnic people. My concern is that these biases could leave certain communities behind if not addressed. What steps are being taken to identify and resolve such biases?
Answer from The Chancellor of the Duchy of Lancaster (Darren Jones)
I thank my hon. Friend for raising that important question. These AI solutions are only as good as their ability to serve the public fairly; equality should be built in at the start. The AI Safety Institute and officials at the Department for Science, Innovation and Technology work across Government to ensure that those values and ethics are built into programmes as they are developed by the Government.
Not addressed: The answer acknowledges the issue but does not provide specific measures or timelines for identifying and resolving biases in AI.
Was this summary accurate? Tell us
How accurate was this summary? (1 = poor, 5 = spot on)