104
votes
Fujitsu bugs that sent innocent people to prison were known “from the start” but concealed from lawyers and judges
Link information
This data is scraped automatically and may be incorrect.
- Authors
- Jon Brodkin
- Published
- Jan 19 2024
- Word count
- 756 words
Holy Hell.
This is the first time in a month I can say that I'm genuinely shocked by a piece of news.
Let this be a lesson to all whom utter "but I have nothing to hide, why should I care."
Because sometimes the system fucks up to the tune of 1,000ish innocent people being convicted.
So horrifying!!
Although I'm not sure that 1,000 people went to jail. I went to this article for more info:
https://www.bbc.com/news/business-67993493
And it's still not clear how many went to jail.
(It's still horrible though even if not all of the people affected went to jail. Just being wrongfully accused is bad enough, that could already lead to ruined finances, relationships, etc.)
The statement could be with a minor edit be stated as "near 1,000 people wrongfully prosecuted and convicted" seemingly, since it's unclear the results of those convictions if they were all jail time or possibly some other punishments.
How does someone even really successfully fight buggy software wrongfully incriminating them? Clearly in this case it seems there were a variety of factors as to why so many weren't able to fight it, but more so I'm wondering just in general what someone can realistically do or what courts will realistically allow for or judges/juries etc. because someone can accuse any software of being buggy, but proving it can be a different story.
English courts say that computers are operating correctly unless there is evidence otherwise. This puts the burden of proof that the computer was malfunctioning on the defence. It takes the burden of proof that the computer was operating correctly away from the prosecution. One simple change would be to remove this, so prosecutions had to show that the software wasn't buggy.
In this case a group of postmasters brought a case to show that the software was buggy, the judge agreed, and that triggered a bunch of reviews of convictions.
This link gives some of the history about how this rule came into existence:
https://www.benthamsgaze.org/2022/06/30/the-legal-rule-that-computers-are-presumed-to-be-operating-correctly-unforeseen-and-unjust-consequences/
As a software developer who lives in England, this is news to me, and all I can say is what the fuck?!
Sure, I understand the need to avoid every single court case being derailed by the defendant demanding a review of a million lines of proprietary code in some unrelated system, but I'd say the fact that none of us can manage a week without running into some level of bug-related inconvenience means that a reversed burden of proof doesn't pass the smell test either. Those day to day bugs might often be just a minor issue, but they present overwhelming evidence that complex technical systems are almost always imperfect for one reason or another.
The suggestions towards the end of that link sound eminently sensible, and pretty much like what I would've described as the "reasonable person" standard I'd assumed was already applied to computers just as it is to many other things in law.
Wow I did not realize it was codified into English law that way. That probably explains a lot behind how this happened for so long, over so many years, before these convictions were able to be effectively overturned years later.
The proposal to address seems fairly reasonable. One reason the question even came to mind before on how to even challenge it is that a lot of software is proprietary, niche, possibly expensive, controlled licensing in some cases, and more now than ever, in the cloud. It makes sense that there should be some kind of disclosure requirements from the side relying on the software/systems because the people in control of those systems are really the ones controlling all the information.
It seems to me that regardless of the bad computer systems, they should be able to follow the money.
If the computer said I stole $10,000 and I didn't, then it must have ended up somewhere else. There must be a source of the money. If it's inventing thousands of dollars out of nowhere then there should be other discrepancies. If the system invents $5000 of income from selling stamps or something, there should be a record of $5000 worth of stamps entering and leaving inventory.
If the system is so fucked that they can't even show where anything is coming or going, then hiw could you ever use it to prosecute someone.
Reading into the case more it really looks like a perfect storm of legal fuckery, where the post office didn't have to prove much, could provide its own employees as expert witnesses and couldn't be forced to turn over records or other exculpatory evidence.
Of course, this requires both work and interest in actual truth and justice on the part of the post office's inspectors. The bonuses they received for "detecting fraud" didn't exactly incentivize them to care about the latter, and without the latter why put in extra work for an outcome that doesn't matter to you?
Incorrect; the appeals, both criminal and civil, began back in 2019. The TV drama is a result of those appeals.
The only party mentioned here that only started paying attention after the TV drama is the mainstream media. The Post Office, Fujitsu, and Government have been "paying attention" as evidenced by the appeals over the last 4 years preceding the television program. The attention that is deserved is the righting of injustice, which has been happening.
It shouldn't have ever happened and the parties that facilitated it should be facing equal punishment for what they wrought on others, but it has been getting deserved attention well before the show.
I haven’t watched the TV show yet, but the book The Great Post Office Scandal by Nick Wallis is a detailed, heartbreaking, and aggravating account of the full story here for anyone interested.
Update: I have now watched Mr Bates vs The Post Office -- the dramatized miniseries based off of this.
For anyone who liked the miniseries, I'd recommend the book. The miniseries does a great job of telling the human side of the story, but its brisk 4 episodes don't go into much technical detail at all.
In that regard, I wish it were longer. The narrative spans decades and hundreds of people! There's so much to show! That said, the series has a genuinely incredible economy to it in that it's able to capture such a long period of time, so many characters, and many different angles and legal matters all in a very short runtime. Not a single second is wasted and every scene is pivotal, but there's hardly time to give anything legs.
The book had the opposite problem: I wanted it to be shorter! It is over-detailed, but I can't deny that's good practice in a situation this large and complex.
One thing I think the show glossed over was the idea that bugs in the Horizon system were genuinely quite significant. The show focuses more on the cover-up angle and the "big organization vs. the little guys" conflict, and it makes it seem like the Post Office was intentionally defrauding the subpostmasters via remote access. The book instead details that the Horizon system was very buggy and architecturally unsound from the very beginning. With the show you can kind of read that between the lines (the balance doubling in the first episode, why Fujitsu would have people working in accounts after hours, etc.), but it's never explicitly laid out.
I know we've got a lot of technically minded people here, so if anyone's wanting a deeper dive into the specifics of the computer side of things, then the book is a great resource.
Regardless, I'd recommend either the book or the show to anyone interested. This is a pretty significant story that deserves more scrutiny. I read the book a while ago so a lot of the specific details have fallen out of my memory, but one that stuck with me was the story of Seema Misra. She was a subpostmaster who was jailed while pregnant with her second child and whose husband experienced racially motivated physical assault while she was in prison -- all while neither she nor he had done anything wrong.
And she is only one of 200+ people that went to prison because of this.
Thank you
Off-topic, but that is a very quick way for me to understand why English speakers are so adamant about degendering language (and why people sometimes fail to understand that this is not as much of an issue in every language).
I wonder, is this just a English as in England thing?
In the US, we would absolutely just say “postmasters” to refer to both male and female. I don’t even know if “postmistress” is a word in American English. To me it conjures up images of some sort of fetish.
Mistress is the female variant of master and the UK loves to put master in titles so when females were allowed at those positions their title became mistress.
It's also the female lover of a married man and I feel like that meaning has been the more popular use in the USA, seeing as how there are no nobility we have to be subservient to and call master or mistress. The last common usage outside the porn industry was probably among slave owners.
Anyway, master is losing it's gendered meaning in the US. Look it up and you see people asking about whether master is genderless. I bet it soon will be recognized as such.
well there's one very salient context in which people were subservient to others and called them master or mistress... but suffice it to say it's not one most people want to evoke in modern day speech.
I do recall seeing webmistress a few times back in the day. But then even webmaster isn't super common anymore.
Postmistress appears in a bunch of US dictionaries, but it's unclear to me whether they're referring to US usage or to non-US usage.
Here are two examples:
From American Heritage Dictionary 4th Ed. (En-En)
From Merriam-Webster's Collegiate 11 (En-En)
Dictionaries are generally trying to be pretty comprehensive, so it's unlikely they'd remove the word if it falls out of use, especially if it only does so regionally. What you'd want here is to consult corpora to see how often these words are used among different speaker populations. I'll update this comment with results if I'm able to find them.
Yes, and I forgot to include the usage note for -ess in American Heritage. That says:
Oh, found this on USPS' website:
https://about.usps.com/who-we-are/postal-history/women-postmasters.pdf
Indeed. I'm approaching this from a US perspective, but we used to have aviator and aviatrix and a wide variety of other gendered titles for jobs