Hacker News

>”Free to ingest and make someones crimes a permanent part of AI datasets resulting in forever-convictions? No thanks.”

1000x this. It’s one thing to have a felony for manslaughter. It’s another to have a felony for drug possession. In either case, if enough time has passed, and they have shown that they are reformed (long employment, life events, etc) then I think it should be removed from consideration. Not expunged or removed from record, just removed from any decision making. The timeline for this can be based on severity with things like rape and murder never expiring from consideration.

There needs to be a statute of limitations just like there is for reporting the crimes.

What I’m saying is, if you were stupid after your 18th birthday and caught a charge peeing on a cop car while publicly intoxicated, I don’t think that should be a factor when your 45 applying for a job after going to college, having a family, having a 20 year career, etc.

> Free to ingest and make someones crimes a permanent part of AI datasets resulting in forever-convictions?

You're conflating two distinct issues - access to information, and making bad decisions based on that information. Blocking access to the information is the wrong way to deal with this problem.

> a teenager shouldn't have their permanent AI profile become "shoplifter" because they did a crime at 15 yo that would otherwise have been expunged after a few years.

This would be a perfect example of something which should be made open after a delay. If the information is expunged before the delay, there's nothing to make open.

"court records are public forever" and "records of crimes expunged after X years" are incompatible.

Instead, we should make it illegal to discriminate based on criminal conviction history. Just like it is currently illegal to discriminate based on race or religion. That data should not be illegal to know, but illegal to use to make most decisions relating to that person.

The actions of the government should always be publicly observable. This is what keeps it accountable. The fear that a person might be unfairly treated due to a long past indiscretion does not outweigh the public's right to observe and hold the government to account.

Alternatively consider that you are assuming the worst behavior of the public and the best behavior of the government if you support this and it should be obvious the dangerous position this creates.

Thanks, it’s super refreshing to hear this take. I fear where we are headed.

I robbed a drug dealer some odd 15 years ago while strung out. No excuses, but I paid my debt (4-11 years in state max, did min) yet I still feel like a have this weight I can’t shake.

I have worked for almost the whole time, am no longer on parole or probation. Paid all fines. I honestly felt terrible for what I did.

At the time I had a promising career and a secret clearance. I still work in tech as a 1099 making way less than I should. But that is life.

What does a background check matter when the first 20 links on Google is about me committing a robbery with a gun?

Edit: mine is an extreme and violent case. But I humbly believe, to my benefit surely, that once I paid all debts it should be done. That is what the 8+ years of parole/probation/counseling was for.

What we do here in sweden is that you can ask the courts for any court document (unless it is confidential for some reason).

But the courts are allowed to do it conditionally, so a common condition if you ask for a lot of cases is to condition it to redact any PII before making the data searchable. Having the effect that people that actually care and know what to look for, can find information. But you can't randomly just search for someone and see what you get.

There is also a second registry separate from the courts that used to keep track of people that have been convicted during the last n years that is used for backgrounds checks etc.

Fully agree. The AI companies have broken the basic pacts of public open data. Their ignoring of robots.txt files is but one example of their lack of regard. With the open commons being quickly pillaged we’ll end up in a “community member access only model”. A shift from grab any books here you like just get them back in a month; to you’ll need to register as a library member before you can borrow. I see that’s where we’ll end up. Public blogs and websites will suffer and respond first is my prediction.

The names of minors should never be released in public (with a handful of exceptions).

But why shouldn't a 19 year old shoplifter have that on their public record? Would you prevent newspapers from reporting on it, or stop users posting about it on public forums?

A thing can't simultaneously be public and not. There is no license to do research nor should there be, so if researchers can get it then anyone can.

If it's not supposed to be public then don't publish it. If it's supposed to be public then stop trying to restrict it.

Totally agree. And it goes beyond criminal history. Just because I choose to make a dataset publicly available doesn't mean I want some AI memorizing it and using it to generate profit.

Records of cases involving children are already excluded so that's not a relevant risk.

Between not delivering the data to AI companies, and barring it altogether is a fair distance. As far as I know, the MoJ is in talks with openAI themselves (https://www.ukauthority.com/articles/ministry-of-justice-rea...).

AI isn't the problem here. Once something goes on the internet it lives forever (or should be treated as such). So has it always been.

If something is expungable it probably shouldn't be public record. Otherwise it should be open and scrapable and ingested by both search engines and AI.

If you commit a crime, get caught, and that makes people trust you less in the future, that's just the natural consequences of your own actions.

No, I don't think if you shoplift as a teenager and get caught, charged, and convicted that automatically makes you a shoplifter for the rest of your life, but you also don't just get to wave a magic wand and make everyone forget you did what you did. You need to demonstrate you've changed and rebuild trust through your actions, and it's up to each individual person to decide whether they're convinced your trustworthy, not some government official with a delete button.

Can you explain your reasoning about “forever convictions”, and for full disclosure, do you have a conviction and are thereby biased?

Additionally, do you want a special class of privileged people, like a priestly class, who can interpret the data/bible for the peasantry? That mentality seems the same as that behind the old Latin bibles and Latin mass that people were abused to attend, even though they had no idea what was being said.

So who would you bequeath the privileges of doing “research”?Only the true believers who believe what you believe so you wouldn’t have to be confronted with contradictions?

And how would you prevent data exfiltration? Would you have your authorized “researchers” maybe go to a building, we can call it the Ministry of Truth, where they would have access to the information through telescreen terminals like how the DOJ is controlling the Epstein Files and then monitoring what the Congressmen were searching for? Think we would have discovered all we have discovered if only the Epstein people that rule the country had access to the Epstein files?

Yes, convictions are permanent records of one’s offenses against society, especially the egregious offenses we call felonies on the USA.

Should I as someone looking for a CFO or just an accountant not have the right that to know that someone was convicted of financial crimes, which is usually long precipitated by other transgressions and things like “mistakes” everyone knows weren’t mistakes? How would any professional association limit certification if that information is not accessible? So Madoff should Ave been able to get out and continue being involved in finances and investments?

Please explain

Names and other PII can be replaced with aliases in bulk data, unsealed after ID verification on specific requests and within quotas. It’s not a big problem.

>Free to ingest and make someones crimes a permanent part of AI datasets resulting in forever-convictions? No thanks.

Is this the UK thing where PII is part of the released dataset? I know that Ukrainian rulings are all public, but the PII is redacted, so you can train your AI on largely anonymized rulings.

I think it should also be against GDPR to process sensitive PII like health records and criminal convictions without consent, but once it hits the public record, it's free to use.

Before you even get to the "AI dataset ... forever-conviction" or copyright issues, you need to address AI's propensity to hallucinate when fed legal data and questions.

> a teenager shouldn't have their permanent AI profile become "shoplifter" because they did a crime at 15 yo that would otherwise have been expunged after a few years.

On the other hand, perpetrating crime is a GREAT predictor of perpetrating more crime -- in general most crime is perpetrated by past perps. Why should this info not be available to help others avoid troublemakers?

https://bjs.ojp.gov/library/publications/returning-prison-0

https://www.prisonpolicy.org/graphs/sex_offense_recidivism_2...

https://usafacts.org/articles/how-common-is-it-for-released-...

https://pmc.ncbi.nlm.nih.gov/articles/PMC3969807/

https://ciceroinstitute.org/research/the-case-for-incarcerat...

https://bjs.ojp.gov/topics/recidivism-and-reentry

I know some countries that emit a "certificate of no judicial history", even when the citizen has so, if they ended the jail time

I think this is wrong, it should be reported entirely at least for 5 years after the fact happened

No, public doesn't mean access should be limited to academics of acceptable political alignment, it means open to the public: everybody.

That is the entire point of having courts, since the time of Hammurabi. Otherwise it's back to the clan system, where justice is made by avenging blood.

Making and using any "profiles" of people is an entirely different thing than having court rulings accessible to the public.

> a teenager shouldn't have their permanent AI profile become "shoplifter" because they did a crime at 15 yo that would otherwise have been expunged after a few years.

The idea that society is required to forget crime is pretty toxic honestly.

Im sorry but that's the equivalent of "I believe in free speech but not the right to hate speech". Its either free or not

We should remember that local journalism has been dead for a decade in most of the UK, largely due to social media.

Any tool like this that can help important stories be told, by improving journalist access to data and making the process more efficient, must be a good thing.

Then it would be even worse if this ends up affecting post-case information.

No. Open is open. Beyond DDoS protections, there should be no limits.

If load on the server is a concern, make the whole database available as a torrent. People who run scrapers tend to prefer that anyway.

This isn't someone's hobby project run from a $5 VPS - they can afford to serve 10k qps of readonly data if needed, and it would cost far less than the salary of 1 staff member.

The issue with that is people can then flood everything with huge piles of documents, which is bad enough if it's all clean OCR'd digital data that you can quickly download in its entirety, but if you're stuck having to wait between downloading documents, you'll never find out what they don't want you to find out.

It's like having you search through sand, it's bad enough while you can use a sift, but then they tell you that you can only use your bare hands, and your search efforts are made useless.

This is not a new tactic btw and pretty relevant to recent events...

Systems running core government functions should be set up to be able to efficiently execute their functions at scale, so I'd say it should only restrict extreme load, ie DoS attacks

If the rate limit is reasonable (allows full download of the entire set of data within a feasible time-frame), that could be acceptable. Otherwise, no.

Without numbers this sounds made up

Totally agreed! This is yet another example of reduced friction due to improved technology breaking a previously functional system without really changing the qualities it had before. I don't understand why this isn't obvious to more people. It's been said that "quantity has a quality all its own", and this is even more true when that quantity approaches infinity.

Yes, license plates are public, and yes, a police officer could have watched to see whether or not a suspect vehicle went past. No, that does not mean that it's the same thing to put up ALPRs and monitor the travel activity of every car in the country. Yes, court records should be public, no, that doesn't mean an automatic process is the same as a human process.

I don't want to just default to the idea that the way society was organized when I was a young person is the way it should be organized forever, but the capacity for access and analysis when various laws were passed and rights were agreed upon are completely different from the capacity for access and analysis with a computer.

Yep, these are commonly sealed records and having worked with family law lawyers there are things that happened to the victims that should never be unsealed.

I like the part where you created a system where if someone has enough resources they can just alter the judicial record.

The problem with "cleaning the data" is it sometimes strips away so much context as to give you a misleading impression. Rory Stewart once said it took him 40 hours to fully understand a piece of legislation he was voting on, yet was expected by the whips to vote on multiple pieces of legislation every week, but most people wave an MP's voting record around like they 100% understood and agree with everything they voted on, despite it being mathematically impossible. If they'd voted differently would it have changed the outcome? Was it even a binding motion? Most of the real debate in the UK Parliament happens beforehand anyway and the government will withdraw any votes they know they're going to lose before it even gets into the chamber so the real rebellions don't even get recorded on theyworkforyou.

Yeah, this is a common problem with these kinds of issues. Data can be open, data can be available, but if it's not easily exploitable and parseable by anyone there will always be a third party that would do so for a premium. Data has to be ready for consumption when shared by the state. That's why one of my very strong objections that we divide datasets by year on a continuous registry. It should be one dataset with a single column that tells you the year the row corresponds to.

Hansard is a great example, because originally parliament did not allow publication of its proceedings; to get round this "Several editors used the device of reporting on parliamentary debates under the veil of debates of fictitious societies or bodies. The names under which parliamentary debates were published include Proceedings of the Lower Room of the Robin Hood Society and Debates of the Senate of Magna Lilliputia."

Well said

I may have misread: https://www.yorkshirepost.co.uk/news/courts/government-suspe...

> “We are also working on providing a new licensing arrangement which will allow third parties to apply to use our data. We will provide more information on this in the coming weeks.

The company in question had a direct relationship with HM Courts & Tribunals Service, and disputes that they sold/distributed any data to any 'AI third party' - says what they actually did was to hire AI-focussed contractors to build some new tool/feature for the platform.

Information is either public or it is not.

If it is public, it will be scraped, AI companies are irrelevant here.

If information is truly sensitive, do not make it public, and that's completely fine. This might have been the case here.

Absolutely fucking crazy that you call accurately describing the reality of AI scraping "absolutely fucking crazy" while at the same time going "who cares?" on attacks against journalism and free speech.

>Oh no, some musician died, PASS THE NATIONAL SECURITY ACT, LOCK DOWN ALL INFORMATION ABOUT CRIMINALS, JAIL JOURNALISTS!!!!

The government provided data to a private company. The private company sold resold access to a third party for AI ingestion. it's a plain case of tough titties to the private company.

That said I don't know why the hell the service concerned isn't provided by the government itself.

When there is a risk of feeding sensitive data to the AI giants the first reaction should be to pull the plug. I'm impressed the government acted quickly and decisively for once. Maybe the company involved will think twice before entering an agreement with an AI company. Notice in the whole rant it is never mentioned which AI giant they were feeding.

Everything aligns with internet conspiracy theories if you try hard enough.

Should we always send the source of truth to a private company to have it repackaged and presented in a more inoffensive manner?

You can get the daily court listings for free online. I would assume this service just provided a ux-friendly front end,etc. which in the clip the minister says they will be replacing.

Yes, predictably…

Why? Immigrants cannot commit crimes?

This kind of logic does more disservice than people realize. You can combat bigotry towards immigrants (issue #1), without covering up for criminal immigrants (issue #2) in fear of increase of issue #1 among the natives. It only brings up more resentment and bigotry.

Actually that’s what data and the preponderance of victims allege: an intersection of immigration and policing which interlocked to systematically deprioritize the investigation into abuse of working-class white girls by an over represented ethnic group.

In the local data that the audit examined from three police forces, they identified clear evidence of “over-representation among suspects of Asian and Pakistani-heritage men”.

It’s unfortunate to watch people and entire countries twist themselves in logic pretzels to avoid ever suggesting that immigration has no ills, and we’re just being polite here about it.

https://www.aljazeera.com/news/2025/6/17/what-is-the-casey-r...

https://celina101.substack.com/p/the-uks-rape-gang-inquiry

This database exposed half a million weekend cases which were heard with zero press notification. Many grooming gang trials were heard this way. The database is being deleted weeks before the national inquiry into the grooming gang cover up begins, and the official reason for deleting the data is nonsensical.

wake up its 2026 even obama and clinton are saying more or less that now about illegal immigration

Why? Grooming gangs went from being a far right conspiracy theory to "there was a coverup" being official government position in a week.

Can you explain your reasoning? Horrible crimes committed by foreign men against native children were covered up for political reasons in the UK. This is common knowledge.

This presumably also falls under the Data (Use and Access) Act 2025 which forbids this kind of citizen data being relayed to third parties without permission. The company don't have a leg to stand on here, which is why it is basing its public appeals now on the impact to its users (journalists). But no company has a right to flout data protection regulations or its agreed conditions of use without serious consequences. Since the data has already been passed on, the breach itself can't be fixed, so it is totally proportionate to order the service to be closed and its data deleted. Frankly, fuck companies with the arrogance to behave this way - cheating agreements and responsibilities in order to make more money, and then expecting indulgence because of the uniqueness of their service.

I've had to deal with the ICO over some pretty minor local government stuff, the bar for reporting to them is very low. I smell a coverup.

The poster linkd to the story they 'could think of', not one that may be upcoming. My guess is on a nonce-case, and the royals are involved.

And how exactly is eliminating a third party search tool for efficiently searching lots of obscure magistrates court proceedings going to stop journalists from paying attention to a spicy court case linked to foreign agents and the PM?

5 Ukrainians. People have traced what some of them were doing professionally when the PM would've been living there. It could be nothing, but we need transparency.

In the UK government records are generally covered by Crown Copyright (which is its own slightly more restrictive weird thing) rather than in the public domain. I haven't checked to see what the status of the court listings are, but the default is very different to the US.

They are non-propagated/effectively hidden.

If you don't "know about them from another source" you can't effectively find/access the information and you might not even know that there is something you really should know about.

The service bridged the gap by providing a feed about what is potentially relevant for you depending on your filters etc.

This mean with the change:

- a lot of research/statistics are impossible to do/create

- journalists are prone to only learning about potentially very relevant cases happening, when it's they are already over and no one was there to cover it

I kept digging and reached the service https://www.courtserve.net. Seems like a windows application (old school one) that receives the data, but I need more time to explore there. They've been working with MoJ for 20 years (their claim). Initially I thought they have people at the courts live reporting but that's a bit of a stretch...