Ben Evans: “A long-running story: in 1999 the UK Post office deployed a buggy software system to its branches that led to almost a thousand people being wrongly accused and convicted of theft, and this is only now being unwound. This feeds into the current ‘AI Ethics’ conversation: there is nothing new about bad software causing harm. Fixing this is partly an engineering issue, but it’s also a process, culture and awareness issue. There will always be mistakes, and the real failure was a refusal to acknowledge that and ask questions.” Link