Twenty-seven years ago, the Law Commission made a decision to trust computers. They reviewed the rules regarding hearsay evidence and concluded that a presumption of proper functioning could be applied to evidence generated by computers. The government agreed and repealed the requirement for showing that the computer system was working properly.
However, specialists in IT law now believe that the Law Commission’s analysis was incorrect. Dr. Sam De Silva, a partner at CMS and former chair of the Law Society’s Technology and Law Committee, stated that the commission misunderstood software, hardware, and IT systems. He believes that their interpretation led to the repeal of the requirement.
The tipping point came in 2019 with the Bates v the Post Office judgment, which revealed that innocent people had been wrongly imprisoned due to presumptions about the reliability of the Horizon computer system. Barrister Stephen Mason and other legal experts had been warning about the dangers of presuming that “the computer is always right” for decades.
In November 2020, a group of nine experts, led by Paul Marshall of Cornerstone Barristers, warned the government about the dangers of this presumption. They emphasized that assuming a computer works correctly is unrealistic and unsafe. They also highlighted that computer errors are not always immediately detectable or caused by incorrect data entry.
The group made several recommendations, including strengthening obligations for disclosure and considering all available evidence when assessing the reliability of computer-derived evidence. However, it appears that no action has been taken on their report.
Even before the findings of the Post Office Inquiry, it was evident that a new approach to handling computer-derived evidence might be necessary. The rise of “large language model” artificial intelligence software has further exacerbated this issue. Some experts suggest subjecting IT designers to ethical conduct standards and requiring registration of programmers for “serious systems.”
In 2018, Sir Geoffrey Vos predicted that computer-generated evidence would greatly reduce court disputes over primary facts. However, it seems that this prediction was premature. The handling of computer-derived evidence continues to be a topic of concern and debate among legal experts and IT specialists.