This subject has a lot of relevance to the Post office scandal in the UK, where local post-office franchisees (“sub-postmasters”) were wrongly prosecuted and convicted of fraud based on software bugs. The whole story is a lot longer than that, and ITV recently done a decent job dramatising part of it and there is currently an inquiry going on.
Changing the rules to have courts interrogate computer evidence would be positive but I don’t agree that this change to the rules “would probably have avoided the disastrous repeated miscarriages of justice over the past 20 years” because a lot more went wrong the Post Office case than just the Horizon software.
The Post Office, acting as prosecutor (!), failed to make essential disclosures and in many cases intentionally prosecuted people they suspected of being innocent and also left people convicted who they knew were innocent based on later information we know they had. It might even be true that the current rules would have worked if the Post Office had disclosed the bugs - as they were obliged to (but did not).
It is really hard to win a case when they other side have custody of the evidence and are refusing to disclose. Changing the rules on computer evidence would be good but I think it’s not right to claim that it would have helped the sub-postmasters. The Horizon miscarriage of justice is centrally about disclosure and not about computers.
People are getting blindsided by the presence of a computer in the Post Office case. The actual problem has nothing to do with the computer. The Post Office knew the computers had problems but they refused to tell anyone. Had they told the defense what they knew about problems with the computers, no one would have been convicted or even prosecuted most likely.
In the US the prosecution must provide all exculpatory (even just mitigating) evidence. Even if the defense doesn’t know about it, it must be offered (although of course not presented, that’s up to the defense). This is the Brady rule and it dates back to a 1963 case (Brady v. Maryland) where the Supreme Court created the idea. To entrench that principle, the same idea exists in many state laws (California even has personal liability for prosecutors that don’t follow disclosure rules) and in the ABA model rules. The federal situation is kind of a mess.
I know far less about the UK legal system, but my understanding is that this idea is very new, isn’t as clearly articulated and applied, and only dates back to the late 90s. Instead of a single rule, there’s a patchwork of different laws that require disclosure, but the Post Office because it had its own prosecution arm did not have any rules requiring disclosure until the late 2010s. And it’s unclear to me if what they have now is effective.
It’s really a case where having a Supreme Court to put its foot down is a good thing (at least when that court sides with the public interest, unlike what we have today in the US). While having a mess of laws results in well.. the Horizon case.
But none of this has anything to do with technology.
There is a High Court of Justice in the UK, which has supervisory authority over all other courts.
The involvement of a computer does actually have some relevance here. When a human witness perjures themselves before the courts, we may determine responsibility. Did they lie deliberately? Were they under some pressure to conceal the truth?
A computer is meant to be unbiased, processing information with neither fear nor favour. And the question of criminal and civil responsibility becomes much murkier with mechanically-produced evidence.
There is a High Court of Justice in the UK, which has supervisory authority over all other courts.
Not at all. It’s a lofty name, but they’re fairly low down. The US equivalent would be something like the District Courts. The High Court of Justice mostly deals with more serious civil cases. See this diagram.
Just like in the US for District Courts, above the High Court in the UK is an Appeals Court and then the UK Supreme Court.
No court has the power of the US Supreme Court in the British system. The UK Supreme Court cannot create a new right and say bind the Post Office to it against an act of Parliament that says otherwise. In the UK there’s the notion of Parliamentary sovereignty/supremacy.
For example, the UK Supreme Court just ruled that the UK cannot deport people to Rwanda because it’s unsafe. So Parliament just changed the law today to declare Rwanda safe and bind judges to that. In the US, this would be absurd, in the UK, it’s just how things go.
The computer is totally irrelevant. The defense should have been given all the materials the prosecution had could help them. That alone would have fixed absolutely everything.
I think the argument is that if the presumption were to doubt computer evidence, then the defence would have a stronger argument to demand disclosure, and would have had support from the judge.
The trouble with disclosure is that you can’t demand the release of a document that you aren’t aware exists. There is a duty on the prosecutor to provide relevant documents. In the case of the Post Office they failed in that duty and it means that even under the current rules, the trials were improper.
Of course, it is ridiculous as we all know to assume that computer software is working correctly…but I think not actually central to that case
You can define “largest” in many ways. I think that if you take into account the number of sub-postmasters affected, the number of judicial processes etc, it might qualify. Other famous miscarriages, even if they affect people in more sever ways, like jail time, don’t affect as many people.
I’m sure there’s a utilitarian formula for this somewhere…
Looking at Locard’s development of the concept of forensic evidence, later referred to as “a silent witness”, the computer is treated in many cases (and, I will argue, usually rightly so) to be a reliable record of fact, a “silent witness”, rather than a true witness - a person whose beliefs and biases may contaminate their testimony before the court.
In the early days of computer evidence, it was necessary for a computer operator to testify that the computer evidence being produced was the product of a computer system working as designed, that it had been verified to be operating correctly (for instance, a test programme being run every morning with a known output to be computed) … with the complexity of modern computer systems, and the intricacy with which modern bad actors can sabotage a system, this is no longer viable.
The question from the legal perspective becomes - how much do we trust the computer, and can one side argue to have computer evidence excluded simply because no human is able to speak for its veracity?
Another question from the legal perspective becomes one of liability - if Post Office systems have effectively perjured themselves and this perjury has resulted in wrongful convictions, as well as economic damages such as loss of income et c.; who bears the responsibility - both the criminal responsibility for perjury, and the civil responsibility for the consequences of that perjury against people who were denied the presumption of innocence due to faulty computer testimony?
Kelman’s “The Computer in Court” is an excellent introductory read regarding potential problems in computer evidence, as well as procedural handling of the admissibility of mechanically produced evidence, but the problem space of computer testimony is only going to get more complex as time goes on.
This subject has a lot of relevance to the Post office scandal in the UK, where local post-office franchisees (“sub-postmasters”) were wrongly prosecuted and convicted of fraud based on software bugs. The whole story is a lot longer than that, and ITV recently done a decent job dramatising part of it and there is currently an inquiry going on.
Changing the rules to have courts interrogate computer evidence would be positive but I don’t agree that this change to the rules “would probably have avoided the disastrous repeated miscarriages of justice over the past 20 years” because a lot more went wrong the Post Office case than just the Horizon software.
The Post Office, acting as prosecutor (!), failed to make essential disclosures and in many cases intentionally prosecuted people they suspected of being innocent and also left people convicted who they knew were innocent based on later information we know they had. It might even be true that the current rules would have worked if the Post Office had disclosed the bugs - as they were obliged to (but did not).
It is really hard to win a case when they other side have custody of the evidence and are refusing to disclose. Changing the rules on computer evidence would be good but I think it’s not right to claim that it would have helped the sub-postmasters. The Horizon miscarriage of justice is centrally about disclosure and not about computers.
People are getting blindsided by the presence of a computer in the Post Office case. The actual problem has nothing to do with the computer. The Post Office knew the computers had problems but they refused to tell anyone. Had they told the defense what they knew about problems with the computers, no one would have been convicted or even prosecuted most likely.
In the US the prosecution must provide all exculpatory (even just mitigating) evidence. Even if the defense doesn’t know about it, it must be offered (although of course not presented, that’s up to the defense). This is the Brady rule and it dates back to a 1963 case (Brady v. Maryland) where the Supreme Court created the idea. To entrench that principle, the same idea exists in many state laws (California even has personal liability for prosecutors that don’t follow disclosure rules) and in the ABA model rules. The federal situation is kind of a mess.
I know far less about the UK legal system, but my understanding is that this idea is very new, isn’t as clearly articulated and applied, and only dates back to the late 90s. Instead of a single rule, there’s a patchwork of different laws that require disclosure, but the Post Office because it had its own prosecution arm did not have any rules requiring disclosure until the late 2010s. And it’s unclear to me if what they have now is effective.
It’s really a case where having a Supreme Court to put its foot down is a good thing (at least when that court sides with the public interest, unlike what we have today in the US). While having a mess of laws results in well.. the Horizon case.
But none of this has anything to do with technology.
There is a High Court of Justice in the UK, which has supervisory authority over all other courts.
The involvement of a computer does actually have some relevance here. When a human witness perjures themselves before the courts, we may determine responsibility. Did they lie deliberately? Were they under some pressure to conceal the truth?
A computer is meant to be unbiased, processing information with neither fear nor favour. And the question of criminal and civil responsibility becomes much murkier with mechanically-produced evidence.
Not at all. It’s a lofty name, but they’re fairly low down. The US equivalent would be something like the District Courts. The High Court of Justice mostly deals with more serious civil cases. See this diagram.
https://en.wikipedia.org/wiki/Courts_of_England_and_Wales#/media/File:Diagram_of_the_court_system_of_England_and_Wales_2014.svg
Just like in the US for District Courts, above the High Court in the UK is an Appeals Court and then the UK Supreme Court.
No court has the power of the US Supreme Court in the British system. The UK Supreme Court cannot create a new right and say bind the Post Office to it against an act of Parliament that says otherwise. In the UK there’s the notion of Parliamentary sovereignty/supremacy.
For example, the UK Supreme Court just ruled that the UK cannot deport people to Rwanda because it’s unsafe. So Parliament just changed the law today to declare Rwanda safe and bind judges to that. In the US, this would be absurd, in the UK, it’s just how things go.
The computer is totally irrelevant. The defense should have been given all the materials the prosecution had could help them. That alone would have fixed absolutely everything.
I think the argument is that if the presumption were to doubt computer evidence, then the defence would have a stronger argument to demand disclosure, and would have had support from the judge.
The trouble with disclosure is that you can’t demand the release of a document that you aren’t aware exists. There is a duty on the prosecutor to provide relevant documents. In the case of the Post Office they failed in that duty and it means that even under the current rules, the trials were improper.
Of course, it is ridiculous as we all know to assume that computer software is working correctly…but I think not actually central to that case
The article describes this as “what is likely the largest miscarriage of justice in British history.”
I…. have questions.
This is iirc the literal terms the High Court used in their judgement
You can define “largest” in many ways. I think that if you take into account the number of sub-postmasters affected, the number of judicial processes etc, it might qualify. Other famous miscarriages, even if they affect people in more sever ways, like jail time, don’t affect as many people.
I’m sure there’s a utilitarian formula for this somewhere…
I dunno; I can think of a few miscarriages of justice in British history that were much, much worse: https://en.wikipedia.org/wiki/First_Opium_War
Not to forget that the East India Company put the Mughal emperor on trial after the Rebellion, despite them being technically his subjects!
Computer: they did crimes
Judge: seems legit
Looking at Locard’s development of the concept of forensic evidence, later referred to as “a silent witness”, the computer is treated in many cases (and, I will argue, usually rightly so) to be a reliable record of fact, a “silent witness”, rather than a true witness - a person whose beliefs and biases may contaminate their testimony before the court.
In the early days of computer evidence, it was necessary for a computer operator to testify that the computer evidence being produced was the product of a computer system working as designed, that it had been verified to be operating correctly (for instance, a test programme being run every morning with a known output to be computed) … with the complexity of modern computer systems, and the intricacy with which modern bad actors can sabotage a system, this is no longer viable.
The question from the legal perspective becomes - how much do we trust the computer, and can one side argue to have computer evidence excluded simply because no human is able to speak for its veracity?
Another question from the legal perspective becomes one of liability - if Post Office systems have effectively perjured themselves and this perjury has resulted in wrongful convictions, as well as economic damages such as loss of income et c.; who bears the responsibility - both the criminal responsibility for perjury, and the civil responsibility for the consequences of that perjury against people who were denied the presumption of innocence due to faulty computer testimony?
Kelman’s “The Computer in Court” is an excellent introductory read regarding potential problems in computer evidence, as well as procedural handling of the admissibility of mechanically produced evidence, but the problem space of computer testimony is only going to get more complex as time goes on.