https://www.bbc.co.uk/news/articles/cy87076pdw3o
Topical....
History has always come in three keys - history as lived experience; history as analytical study; history as myth. AI will fill our world with more myth takes. ::)
PS - although out of period, I do recommend Paul Cohen's "History in Three Keys", his study of the Boxer uprising period in Chinese history.
This message has no AI content, as far as I am aware :-\
::)
I think it shows a potential weakness in using AI for historical reconstruction - how does the AI judge from its vast data source what is historically accurate and what isn't? How much is it guided by a human and how skilled is the human at making those judgements? With a video game or a piece of TV or film, we might say its OK to compromise reality for the sake of entertainment, spectacle and story. If the idea is to teach, even if the intention is to help visualise what the past was like, there is more of an obligation to stick to the facts or at least reasonable interpretations thereof.
For me, Kim boils the question down to its essence and Anthony expands on the creative aspect of recreating fact-based history. Historians and those writing history must make choices. There is a seemingly inexhaustible array of choosing what to include and what not to include in the writing of history. Writers often have an agenda or particular perspective from which they write. Often, sources are plucked to support this narrative. Hopefully, these choices are made thoughtfully and intelligently. Not something I trust to AI generated histories unless we are interested in reading historical fiction or myth as in Kim's Third Key.
I touch on some of these AI pitfalls in a recent forum post on AI Designed Rules (http://soa.org.uk/sm/index.php?topic=8966.msg114318#msg114318).
I think historians can be a bit over-precious on this sort of thing.
If this gets more people interested in history, then it's a great thing. Yes, it may be inaccurate but what popular history isn't?
The idea that people's gateway into history should be via peer reviewed, academic standard articles is pretty laughable, really.
Quote from: Cantabrigian on February 23, 2025, 04:29:48 PMThe idea that people's gateway into history should be via peer reviewed, academic standard articles is pretty laughable, really.
Which is the reason I've always been a fan of well-written popular history :) Or popular historians on TV. But, if there is no peer review, the obligation to get it right (even if it is condensed or simplifies complexities) puts a great burden on the author. I would apply the same standards to those peddling AI-assisted popular history.
I agree. I like historical dramas and novels but they have to be based on a good solid historical foundation for me personally
My concern about these videos is that the images they produce will become the dominant historical viewpoint for the general population. The producers have a responsibility to stay within the bounds of generally accepted historical research. The producers may present a different interpretation of the event, but they should not ignore or alter the basic knowledge. So, no lava burying Pompeii.
Popular histories can inspire and educate without being a dull, ponderous, peer reviewed piece of writing that can serve as a cure for insomnia. An example of an accurate, well written popular history is Battle Cry of Freedom, The Civil War Era by James M McPherson, a professor of American History at Princeton University. This book won the Pulitzer Prize for history and was a national bestseller that did educate the reader. I regard Professor McPherson's book as an example of what popular history should be.
Quite right too...the danger of 'popular' content without due diligence is misinformation
The problem comes when history is generated with an eye to modern politics and control
This has been a problem in every age, surely? Didn't the Egyptians give a rather different spin to Kadesh than the Hittites?
We're all products of our political environments, and historians are no exception!
8)
Quote from: Keraunos on February 24, 2025, 11:30:03 AMThis has been a problem in every age, surely? Didn't the Egyptians give a rather different spin to Kadesh than the Hittites?
In the past, historians did not necessarily see their primary task as recounting the facts of past events. They may see the moral or religious or regime glorifying aspects as more important. Which is, of course, why we put store by understanding the agendas or simply social conventions of past historians.
This kind of AI is just a giant internet index system that responds to a query with a search and uses its software to mesh the relevant material together in a single presentation. It is utterly dependent on human input and the importance the internet logarithms place on that input. Which means, meh, what's the big deal? I'm waiting for all this excitement over AI to wear off. We got over the hula hoop so I'm hopeful.
PS: I just turned 60 so am entitled to be a grumpy old man. >:(
Quote from: Keraunos on February 24, 2025, 11:30:03 AMThis has been a problem in every age, surely? Didn't the Egyptians give a rather different spin to Kadesh than the Hittites?
That wasn't history, that was news ;)
Quote from: Prufrock on February 24, 2025, 12:06:43 PMWe're all products of our political environments, and historians are no exception!
That is why I still pick up history books written a century ago
Depending on the topic the knowledge can be dated, but the opinions of a previous age are a nice counter to the opinions of a current age.
I looked at my stuff on the Greeks in India and Bactria and discovered my books range from high Victorian adventurism through Pre-War certainty, the fine flowerings of Indian Nationalism, and Late American Colonialism 8)
99% of the population isn't going to read a book on the ACW however well written. So the real choice is between something like this, and absolutely no historical knowledge at all.
Personally I'd go for the former, because it's a gateway to more detailed and more accurate stuff. Bore them the first time, and they'll never come back.
You may think that these videos are inaccurate, but they're nothing compared with what most of the population believes...
Quote from: Cantabrigian on February 24, 2025, 01:37:35 PM99% of the population isn't going to read a book on the ACW however well written. So the real choice is between something like this, and absolutely no historical knowledge at all.
Personally I'd go for the former, because it's a gateway to more detailed and more accurate stuff. Bore them the first time, and they'll never come back.
You may think that these videos are inaccurate, but they're nothing compared with what most of the population believes...
Most people never do the deep dive into any topic to the extent they acquire an accurate understanding of it. I tend to think a little knowledge is dangerous since it creates spurious certitudes (the Roman upper class were absolutely corrupt to the extent of having a room dedicated to throwing up when they over ate), so absolute ignorance is perhaps the better option. At least you know you don't know.
Quote from: Erpingham on February 24, 2025, 12:52:31 PMQuote from: Keraunos on February 24, 2025, 11:30:03 AMThis has been a problem in every age, surely? Didn't the Egyptians give a rather different spin to Kadesh than the Hittites?
In the past, historians did not necessarily see their primary task as recounting the facts of past events. They may see the moral or religious or regime glorifying aspects as more important. Which is, of course, why we put store by understanding the agendas or simply social conventions of past historians.
Which historians do you have in mind, and to what extent did they distort the facts in favour of a moral and religious agenda?
That's a tricky one. I won't be able to say much on classical sources, though the name Suetonius is often mentioned. Plutarch, I believe had something of a moral agenda (though I must admit he always seemed fairly good at his job to me).
Medieval stuff I'm more familiar with. So, the lives of saints tend to be stronger on religious significance than factual reportage. European crusades histories tend to have very pro-crusade, anti-pagan narrative. The legitimacy of monarchs actions are often written of as confirmations of just cause and divine favour.
I'm sure there is plenty more but I think gives a general perspective.
Quote from: Erpingham on February 24, 2025, 02:10:49 PMThat's a tricky one. I won't be able to say much on classical sources, though the name Suetonius is often mentioned. Plutarch, I believe had something of a moral agenda (though I must admit he always seemed fairly good at his job to me).
Medieval stuff I'm more familiar with. So, the lives of saints tend to be stronger on religious significance than factual reportage. European crusades histories tend to have very pro-crusade, anti-pagan narrative. The legitimacy of monarchs actions are often written of as confirmations of just cause and divine favour.
I'm sure there is plenty more but I think gives a general perspective.
I suspect that if one is looking at historians who are trying to be historians most of them can be quite reliable, at least to the extent of not allowing religious or moral considerations to distort the facts as they know them. Suetonius though was more like a modern journalist. Don't let facts get in the way of the spice. ;)
Lives of the saints....depends on who's writing them. Of course the conviction that they're all tosh rests on the assumption that miracles and any supernatural event are impossible, and that is an assumption.
Crusader histories...I'm weak on these. Can one distinguish between reliable and non-reliable ones?
Quote from: Justin Swanton on February 24, 2025, 03:16:37 PMLives of the saints....depends on who's writing them. Of course the conviction that they're all tosh rests on the assumption that miracles and any supernatural event are impossible, and that is an assumption.
To go back to my original suggestion, I think the primary purpose of a saint's life was to glorify the saint (hence we have a special term for this type of writing). I don't think they are useless as history but we must approach them with their purpose in mind.
QuoteCrusader histories...I'm weak on these. Can one distinguish between reliable and non-reliable ones?
Not my subject either (I'm not that interested in the Crusades tbh) but I understand some are more reliable than others. Joinville is supposed to be quite good, as we get a soldier's eye view. But all will see the situation based on their religious background (understandably).
A few of these have popped up on Instagram, they are hilarious
I don't see a vast amount of difference between AI-generated historical scenes (complete with inaccuracies) and, say Ridley Scott-generated historical scenes (complete with inaccuracies).
In the case of the latter, the excuse is always "it's just entertainment" or "nobody would think it was a history lesson" - which are both lame excuses IMHO as accuracy can be just as entertaining, and lots and lots of people think it is a history lesson.
Visualising history is a worthwhile activity - I always think of history on three levels: what happened, why or how did it happen, and what was it like (which is the poor relation). Videos whether made by AI or humans can help with the last of these, but if not as accurate as possible, there's no point. My experience of AI content to date does not fill me with confidence that the accuracy will be high enough to make the exercise worthwhile (the same observation applies to the films of Ridley Scott).
Quote from: RichT on February 24, 2025, 05:43:35 PMVisualising history is a worthwhile activity - I always think of history on three levels: what happened, why or how did it happen, and what was it like (which is the poor relation). Videos whether made by AI or humans can help with the last of these, but if not as accurate as possible, there's no point.
A good summary. Good reconstruction is a way to help people relate to history. As Mike said, we can't expect everyone to dive into reams of peer reviewed text written by people with poor communication skills - they will be put off before they get to the interesting bits. So gateways are important.
As I understood the chap making the videos who was interviewed at our start point, he isn't just asking the AI to make a video and then put it up. He has some hand in the creative process. Like a director working with humans, he needs to balance his intention (art, drama, entertainment, education) and get it into the product. I don't think AI can be praised for the good bits or damned for the rubbish bits really.
And to be fair it's been done for centuries. Ancient authors weren't above embellishments and spin...
Quote from: Justin Swanton on February 24, 2025, 01:46:06 PMI tend to think a little knowledge is dangerous since it creates spurious certitudes (the Roman upper class were absolutely corrupt to the extent of having a room dedicated to throwing up when they over ate), so absolute ignorance is perhaps the better option. At least you know you don't know.
Stepping outside your front door is a dangerous thing - you never know where it may take you.
The problem is that if you want anyone to have a lot of knowledge, then they have to go through a stage of only knowing a little.
Quote from: Cantabrigian on February 25, 2025, 07:57:20 AMQuote from: Justin Swanton on February 24, 2025, 01:46:06 PMI tend to think a little knowledge is dangerous since it creates spurious certitudes (the Roman upper class were absolutely corrupt to the extent of having a room dedicated to throwing up when they over ate), so absolute ignorance is perhaps the better option. At least you know you don't know.
Stepping outside your front door is a dangerous thing - you never know where it may take you.
The problem is that if you want anyone to have a lot of knowledge, then they have to go through a stage of only knowing a little.
The real problem is that to know a thing truly you have to know it in depth and nobody has the time to thoroughly research every topic. That means we rely on summaries done by others for almost everything we think we know. A summary by itself is already inaccurate and one has no way of knowing if the source of the summary is somebody's right thumb or something more substantial. I mean....look at the quincunx! :o ;)
Another problem might be the sheer speed that junk can now be churned out. One computer can create in seconds the sort of tripe that previously would take a film crew, a writer, a handful of actors, and a quarry somewhere to create.
Quote from: Justin Swanton on February 25, 2025, 08:09:48 AMThe real problem is that to know a thing truly you have to know it in depth and nobody has the time to thoroughly research every topic. That means we rely on summaries done by others for almost everything we think we know.
True. Of course, if we go large on this, there are enough of us humans for us to have people who know all sorts of stuff in depth. The trick is to find the trustworthy sources for the information we need.
Quote from: Erpingham on February 25, 2025, 10:33:59 AMThe trick is to find the trustworthy sources for the information we need.
There lies the rub.
A still image rather than a video but illustrating some of the issues of AI illustration
(https://kkrva.se/wp-content/uploads/2025/02/Agincourt.jpg)
I think the commissioner has asked for an image of the Battle of Agincourt in the 19th century Salon style. This would explain the rich colours, anyway. However, a few seconds looking and you start to see some weird stuff.
Most of the men seem to be armed with curved sticks, which are waved or poked at the enemy. in the middle distance, some sticks seem to be in the air, so they must be throwable. In the left foreground, two men seem to passing a curved stick between them. Based on the perspective it must be about 12ft long.
Try to sort out the anatomy of the bodies in the left foreground. I couldn't.
A man in the right foreground has his helmet on backwards
The strange mummy like figure on a horse in the left middle ground seems to have a stick in his left hand, another in his right and with his third hand he appears to hold the horses mane
It is unclear why the participants are fighting in a thick fog
And so it goes. I suspect the curved sticks are because the AI was told to include bows. It could tell what their shape was but had no idea how they were used.
Incidentally, the image comes from the The Swedish Royal Academy of War Sciences, so maybe Agincourt is a bit of a marginal subject for them. :)