• Succession

This very point was proved shortly afterwards when I’d been explaining some other historical point (to do with royal succession in the twelfth century, since you ask). The following day in school, he was publicly shot down by the teacher for repeating this, despite the fact that he and I were right and she was wrong.

If you’re teaching people, you need to get your facts straight. I also wasn’t going to have my son wrongly humiliated in class. So, in I marched to see the Head Teacher and explained why I was so cross. I suggested that, if he didn’t believe me, he could look the subject up on Wikipedia. “They have it right,” I said. “She didn’t.”

• Collision

I can’t think of a website I use more often. I look up the oddest things on it, sometimes for professional reasons. It has a number of advantages over any other source. In no particular order, it is consistently laid out, is free, has no adverts, is comprehensive, is clearly written and it shows its sources.

Above all, it’s crowd-sourced. There are over 260,000 contributors. I spoke to one this week. He confirmed that they don’t take just anyone: you can’t write in and say “I’m a Man City fan who knows a bit about WW1 – give me a log-in.” Wikipedia kind of knows who all the contributors are, although some choose (for reasons of personal safety, such is the site’s influence and thus the threats they receive) to remain anonymous.

This aspect has brought Wikipedia into collision with the government’s Online Safety Act (OSA) which says that major online platforms must verify the identities of their contributors. A recent High Court judgment “dismissed Wikipedia’s claims that a new user verification rule will end up hurting the free digital encyclopedia’s editors”. However, the Judge added that the ruling “should not be seen as giving regulators a “green light to implement a regime that would significantly impede Wikipedia’s operations.”

This seems like the very definition of an equivocal response. It seems clear that, almost as he was speaking, the Judge realised that he was looking at another poorly drafted piece of legislation.

• Differences

Let’s take two different examples. The first is a hypothetical social-media post or a blog which urges that, say, immigrants, refugees or asylum seekers should be persecuted. The second is Wikipedia’s page on the National Front.

The OSA is – rightly, as many would agree – aimed at the first group. Online trolling, cyber-shaming and keyboard vigilantism is everywhere. It’s very easy to spout whatever comes into your head if your user name is xyz999. Even your partner or next-door neighbours might not know what you’re up to. Given that there’s legislation about hate crime – how effective, relevant or useful this is is a separate issue – it’s logical that the state should need to know who’s written this stuff. Otherwise, it’s a bit like trying to enforce traffic regulations if there are no number plates and no driving licences.

Is Wikipedia – which is without doubt “a major online platform” – in the same category? No.

The online blogger is probably one person. No one else can edit what’s been written. Wikipedia is utterly different, and in ways that destroys any relevance it has to this legislation.

Let’s take a closer look at the above-mentioned NF article. It’s nearly 11,000 words long. It has over a hundred references to other Wikipedia pages. It has 408 cited references. It has twenty-seven other published sources cited for further general reading. It has been edited five hundred times since March 2020 alone, seemingly by at least a hundred different people. Any of the other 260,000-odd contributors could also edit it if they wanted.

• Bothered?

This leaves me not at all bothered as to who contributes to a Wikipedia page. The mere fact that so many others can step in and edit acts as a brake on extreme views. This doesn’t apply to the loner in their bedroom who has no such moderating influences.

If you visit a Wikipedia page regularly you may see that it’s changed since the last time. That’s as it should be – life moves on, as do the views of those writing about it.  The blogger in the bedroom – and, for that matter, the once-undisputed academic paper or newspaper article – is often locked in the time it was written. It conveys all the authority of being the last word on the matter but is often not updated to reflect future developments of facts or thoughts. A lot of work is needed to track down why this might be seen as flawed. Wikipedia will often provide this.

Sometimes the most bizarre changes are made to Wikipedia entries. I was partly responsible for one of these.

• Catherine

Many years ago, I was following a cricket match on the BBC’s test Match Special’s website. As is so often the case, the comments became only tangentially related to the match (often the best TMS commentary is when there’s no play to commentate on due to adverse weather). The online discussion had moved onto things said by famous people from a pre-cricket age which could refer to the sport, were you to twist their words enough.

I recalled that when King Charles II was first introduced to his wife-to-be, Catherine of Braganza, he turned to the courtier who had arranged the match and hissed “you have brought me a bat, not a woman.” I mentioned this, you understand, by way of historical reportage and to contribute to the ongoing debate, not to condone his unfeeling comment.

This sparked further examples, none of which I can now recall. Eventually, the rain stopped and play re-started.

The next day, I visited Catherine of Braganza’s Wikipedia page to double-check a fact about her I’d later mentioned. Imagine my surprise when I read (and I’m paraphrasing), “Catherine was even featured on TMS on [whenever the date was] when her appearance, and cover drives, were debated during the Test Match between…”

I went back half an hour later and this had been removed: as it should have been. It was a brief joke: you laugh and move on.

• Reliability

Questions about Wikipedia’s reliability remain and are summarised on this page on (where else) Wikipedia. Accusations include long-term uncorrected errors, the fact that most of the contributors are male, ideological bias, revenge editing and the perpetuation of misinformation cycles.

Much the same charges could be levied at many sources which, being longer established, are perhaps not subject to the same strictures. The above comments, for example, perfectly describe many of our national newspapers.

It also says that “Wikipedia is likely the most important single source used to train generative artificial intelligence (AI) models, such as ChatGPT, for which Wikipedia is valued as a well-curated data set with highly structured formatting.”

It also adds that “the accuracy of AI models depend on the quality of their training data, but these models are also fundamentally unable to cite their original source for their knowledge, thus AI users use Wikipedia knowledge without knowing that Wikipedia is its source.” If this is a problem, then it’s surely one that AI needs to fix. The sources are all there.

• AI

I would feel happier with the intention of the Online Safety Act were it also to demand the provenance of every author of every source that has contributed to the various AI systems. I remember writing some weeks ago about a pro-Russian website which, according to a BBC R4 programme, had added one and a half million articles in 2024. All of these would be used by AI harvesters to create what purports to be The New Truth.

Anyone doing a search on Google will have noticed that an AI overview is often the first result returned. I asked a question last week – about cricket, once again – and clicked on this option out of curiosity. Take it from me, the result was gibberish. Getting a statistical query badly wrong hardly counts as a hate crime but it does evince a worrying lack of fact-checking.

The ideal AI could be defined as something that takes the sum of the world’s knowledge and distils it into an accurate and digestible summary. To a large extent, this is what Wikipedia already does. Moreover, it does it by mainly using humans – although it admitted in May that it “was integrating generative AI into its editing process as a means to help its volunteer and largely unpaid staff of moderators, editors, and patrollers reduce their workload and focus more on quality control.”

I’m not completely comfortable with this as it could be just another feedback loop, with the increased use of AI feeding more AI. However, with about 64 million separate articles on Wikipedia, it’s easy to see why some extra help is needed.

260,000 human contributors is, in these times, perhaps not enough. The current legislation, if rigidly interpreted, seems likely to reduce this number. I’m not sure that this is its intention and I hope that further court rulings will confirm this.

• Another scandal

Last week, I wrote about some of the scandals – ranging from railway accidents to infected blood and from child abuse to Postmaster persecutions – which seem to be both covered up and not quickly resolved. Here’s another one. This concerns, as described in this BBC article, a bug in the HM Courts & Tribunals Service (HMCTS) system that probably resulted in judges in civil, family and tribunal courts having made rulings on cases when evidence was incomplete.

I spoke to one local solicitor, Karen Salmon of Marlborough Law, about this. “It comes as no surprise to most family solicitors who spend a lot of time in the Family Court,” she told me. “Several times I’ve attended court only to be told that certain important documents sent in the required time had not reached the Judge or even appeared on the portal.” The method she’s hit upon is to hand in a paper copy a few days before the trial.

However, this rather negates the purpose of having an online system at all. What happens if your lawyer in a civil or family case trusts the system but your opponent doesn’t? The omission of major documents would, as mentioned above, generally be spotted. More minor ones going missing might not be. These could be important in a finely balanced decision. It would be a real worry if participants assumed that Judges had read documents which they had in fact never even seen.

The HMCTS has, however, “dismissed claims that a technology failure adversely affected cases, stating that an internal investigation found ‘no evidence’ of any impact on outcomes.” Others, including Local Government Lawyer, were more sceptical. The internal investigation certainly doesn’t seem to have been that rigourous.

Equally sceptical is the Bar Chair, Barbara Mills, who has, as reported in The Law Gazette, called for an “urgent and comprehensive investigation into the incident” which she describes as the “latest in a series of issues with the courts’ IT systems which it’s now apparent are affecting the entire justice sector due to lack of investment”.

She also goes further than this one issue, arguing that “The Ministry of Justice should undertake an audit of all its IT systems so that any problems can be proactively identified and addressed rather than sticking plasters being used as and when these issues come to light. Our IT systems must be fit for purpose before they are implemented.” The clear implication is that she feels that this system, and perhaps others, fall short of this minimum requirement.

In short, the HMCTS stands accused of, at best, belittling and, at worst, covering up this problem. Anything that affects the way justice is administered is really serious. If that goes wrong then the last check we have before a judgement is passed and lives are changed goes bango: and, in this case, all because of an IT glitch.

By extension, there’s a wider point to be made. If the HMCTS claim that no cases were adversely affected is true, that suggests that a random selection of evidence is acceptable to arrive at a fair decision. I knew someone who worked in HR years ago who, to cut her workload by 33%, randomly sorted CVs into three random piles and threw one away. She could, perhaps, have picked just one CV and offered that person the job without interview. The difference is only a quantitive one. This seems to be rather like what’s happening here, though with more serious results.

Taking this to its extreme, one could also argue that having no evidence at all is fine. The parties file in, the Judge takes one look at them and delivers a verdict. This is what most of us do in social interactions – apparently we decide in about the first five seconds if we like someone or not. It’s not, however, a good basis for deciding legal cases (though it would certainly be quick): yet that’s the direction of travel that HMCTS’s insouciance seems to be encouraging.

And finally…

• Trump has plans to end the war in Ukraine with a meeting with the Kremlin Crocodile in Alaska this week. Zelensky, who had wanted to be involved, has had to content himself with another global PR offensive. Trump wants to make a deal: Putin doesn’t need to do anything.

As in WW2, the number of casualties – a million killed or injured so far in Ukraine, according to several estimates including this one – don’t signify. For him, and his yes-people, it’s an existential struggle against Fascism. I’ve yet clearly to understand the right-wing and left-wing distinctions between extreme political behaviour. I’ve always distrusted patriotic and nationalistic motivations and Putin’s use of it has done nothing to change my mind.

• The BBC reports that the government has announced that “facial recognition (LFR) vans will be rolled out across seven police forces in England to locate suspects for crimes including sexual offences, violent assaults and homicides.” This has led to a predictable range of concerned reactions, some of them justified.

The trouble is that many of us tend to trust the Police when they’re dealing with a crime that’s affected us (if they can produce a quick response) but rather less so when it comes to more general matters of wider policy, or the ability of the forces to control the baser instincts of some of those they recruit. The job calls for the very highest standards, to which most of them live up to. Some don’t: which tarnishes everything. As a result, many people don’t trust initiatives like this.

• There seems to be an unlikely bromance building up between Foreign Secretary David Lammy and US Vice President JD Vance, which has been slightly soured by the fact that it seems that Lammy at least didn’t have the necessary licence when they recently went fishing together. Predictably, this was blamed on an “administrative error”.

• Eighty years ago last week, on 6 and 9 August, atomic bombs were dropped on Hiroshima and Nagasaki. This article in The Week look back at these seminal and controversial events.

• To change the mood completely, the cricket Hundred continues to delight me, often to extent of getting nothing else done when the women’s and men’s double-headers are available on the BBC. Yes, I know it’s very swish-bash-bosh but I love the on-field insights from the players, the banter in the commentary box and the fact that the whole game can change with a couple of wickets or a flurry of boundaries.

Sport is on many levels very facile but it also tells us something useful about the pursuit of excellence and the power of the completely unexpected. I also love the way that this tournament treats the two genders more or less equally. The two matches have their differences but, taken together, provide a welcome, though, temporary release from the awfulness of more or less everything else. I admit it’s an escape: but I suspect there are worse ones than cricket…

Across the area