 So, welcome to day number two of EMF camp. I think it's day two, yes? Yeah, yeah, day two of EMF camp. Good afternoon. Welcome. You're all seated. You're all relaxed. So without further ado, I would like you to give a very warm welcome to Henry Cook and things that go bump on the internet. Thank you, everybody. Thank you, EMF camp, for having me. And thank you to everybody for turning up. I'll start with a disclaimer. This talk is at something like version 0.8. There are probably bugs. By continuing to sit here, you agree to be bound by the terms of my beta test agreement. If you think I've missed a trick or a particularly good story, please come and find me later. I've sort of barely started scratching the surface of the subject, and it's really fun. It turns out that this is an ambitious topic to try and fit into half an hour. I'll be glossing over a few things. Passing by a few others at speed. I'll assume that everyone is smart and has Google, so if I say something and just leave it, you can look it up later. I've also published a lot of notes online, and I'll put up the links to that at the end. OK, with all of that get said, let's get started. 174 years ago, on May the 24th, 1844, Samuel Morse sent the first long-distance telegraph message across America from Washington DC to Baltimore. Does anyone know what the content of that message was? Someone, a few people. You'll enjoy this then. Pretty dramatic. Now, you can't help but read this as monumental hubris, or at least maybe some sort of premonition on Morse's part concerning the communications revolution he was starting to bring about. But it's also interesting for another reason, because it frames the telegraph as a gift from a higher power, from God to Morse to the rest of us. Morse is merely channeling God's will. Four years after Morse's transmission, 1848 in New York State, two sisters reported hearing knocking, rapping, and banging sounds at night that weren't being made by anyone else in the house. In time, the Fox sisters started to communicate with the entity making these sounds. On March the 31st, Kate, the younger of the two, invited it to repeat the snappings of her fingers. It did. She then asked it to knock out her age and that of her sister Margaret, and that of their older sister Leah, who no longer lived with them. It did this too. 12, 15, 17. Over the next few days, they devised a code which included yes, no, and letters of the alphabet. Then the Fox family fled the house, moved to nearby Rochester. The tapping spirit followed. Kate and Margaret became famous, demonstrating the first paid public seance in November 1849 and going on to successful careers as spirit mediums until their confession in 1888 that it was all a hoax. By that time, though, it was far too late. Spiritualism had well and truly taken hold in a popular British and American imagination. In his book Haunted Media, Geoffrey Scont suggests that the story of the Fox sisters and that of Samuel Morse are far from unrelated. The telegraph, along with its fundamental mechanism for sending messages over long distances, had brought with it a whole new way of thinking about communication with other bodies. Once you've abstracted communication out to something that can happen instantaneously with no idea of who or what is on the other end of the line, then it's easy to take that abstraction one step further and imagine communication with entities on other planes of existence, communicating through far more ethereal means than a simple wire. Now, at this point, I really should mention radio and electronic voice phenomenon, EVP, but to be honest, that would take an entire half hour easily by itself, but that's definitely one to look at later if you're interested and you don't already know about it. 83 years on from Kate and Margaret's tapping spirit, 1931, accounts of a contact with strange creatures or spirits weren't that uncommon. The spiritualist movement, kick-started by the accounts of the sisters Fox have been running for a while now and people had been communing with voices from other planes of existence with some regularity. In an isolated farmhouse on the Isle of Mann, the Irving family, Jim, Maggie, and their 12-year-old daughter, Voyery, reported hearing scratching and rustling noises coming from inside the hollow walls. These noises were apparently made by a trapped mongress, one with yellow fur and hands and feet, who had been born in New Delhi in 1852. They knew this because Jeff, as he called himself, spoke to them. Now, I went to talk a little while ago about Jeff and they showed us some archive footage of Voyery being interviewed. She imitated Jeff's voice, so I'm going to do my best attempt to imitating Voyery, imitating Jeff for you now. I am, but I shan't tell you. I am a freak. I have hands and I have feet, and if you saw me, you'd faint. You'd be petrified, mummified, turned into stone or a pillar of salt. Of course, Jeff couldn't speak at first. He just squealed and thumped inside the walls of the house. However, in time, he started to make gurgling noises like a child, and with encouragement from the isolated Irvings, he picked up English, not least through Jim's daily reading to him from the newspaper. Jeff could also speak the odd phrase of Russian, Spanish, Welsh, Hebrew, Manx and Hindustani. He was a well-traveled mongoose. As he himself is reported to have said, I was brought to England from Egypt by a man named Holland. When I was in India, I lived with a tall man who wore a green turban on his head. Then I lived with a deformed man, a hunchback. As Jeff's confidence grew, he started to explore the island and brought gossip back from what he'd heard while catching rides under the buses. He gained a nest under the rafter's envoys room, which is a family called Jeff's Sanctum, and they left bacon and sausages for him there, which he swiped while they weren't looking. With his confidence grew his notoriety. Jeff's story spread around the island, and in 1932, it was picked up on by the Isle of Man examiner and the Isle of Man weekly times. From there, it was picked up by the mainland press, and before long, the Irving's farmhouse was getting frequent visits from journalists and paranormal investigators. Jeff never appeared to any of them, though. Although he would rattle and scream while they were there, he saved particular ire for one Harry Price and his dictaphone. Is it that spookman Harry Price? Why, I won't speak into it. I'll go and smash his windows. I'll drop a brick on him as he lies in bed. Of course, the most likely explanation is that Jeff was, as sconce puts it, an imaginary companion created by voyeurie, the creative girl's reaction to prolonged isolation and boredom. The story of Jeff adds another layer to the spiritualist legend, though. If we accept, indeed, that Jeff was a hoax, then it was one enabled and accelerated by the communication technologies of the day. Voyeurie's imagination was fed by the information that made it to the farmhouse at Cashion's Gap, through books, newspapers, and radio. This is what makes Jeff more interesting than a usual haunting where the voices tend to bang on about the ethereal mysteries on the other side of the veil. I'll split the atom! I am the fifth dimension! I am the eighth wonder of the world! The story itself spread on the same networks that fed it. The tabloid press knowing a good story when they saw one gave extensive coverage to Jeff and by 1935, he'd made it as far as the Hong Kong Telegraph. The story's so potent that it's continued to propagate down through time to the present day, jumping from newspapers to books to magazines and the internet, and even to a field in Gloucestershire in 2018. At the last DMF camp, I talked about number stations, mysterious shortwave radio stations operated across Europe during the Cold War, mostly broadcasting strings of numbers. Most people think that these were coded communications channels, broadcast by state intelligence agencies, although none have ever fessed up to running one directly. During that talk, I said that the thing that I found interesting about the number stations was the way that they acted as a kind of generator of folktales. And I reckon that you could draw a line back from the stories a number station spawned to campfire ghost stories. Today, I'm talking about my attempts to trace that line, how existing folklore makes its way into how we think about technology and how our technologies themselves generate their own folklore. One of the earliest examples I could find of a technology spawning its own distinct folklore, as opposed to existing ideas about spirits manifesting in a new space, was Titavillus, the printer's devil, who crept into print shops and introduced errors into carefully laid out type while the printers were sleeping. There's also an idea in Japanese... Oh, that thought is in the wrong. 0.8. There's also an idea in Japanese folklore of Tuscumigami, spirits which tools acquire when they get old enough, about 100 years. These spirits are harmless enough, mostly playing whimsical pranks on their owners, unless you annoy them by throwing them away too soon or being wasteful. Galvani's experiments, twitching frog's legs with electricity in the late 18th century, gives us Mary Shelley's Frankenstein, and a few steps later, gives us the killer AI myth in The Terminator, where an artificial child reeks revenge on its creator. One of the first examples I could find in the 20th century are gremlins. They're these mischievous little imps, and they were said to infest aeroplanes, and they had an advanced knowledge of aeronautical engineering, which they used for devoury. It was common for pilots to blame mechanical problems with their aircraft on the gremlins. Some even claimed that the gremlins could communicate them with them on a psychic level, causing their vision to fog or to see mountains which weren't there. Here's an apparent account by LW, a B-17 flying fortress pilot during World War II. So I am very aware of my surroundings, and as I go higher, I notice an unusual sound coming from the engine. The instruments went nuts. I look at my right, and I see an entity staring at me. Then I look at the aircraft's nose, and there it is, another one hanging in there, dancing lizards, but I was perfectly fine. My senses were in good shape, but the weird things were still there looking at me. They kept going at it, pounding the plane with all their might. They appeared to be laughing, with their big mouths open, looking at me, hitting the plane with their long arms, trying to pull stuff. I have no doubt in my mind that they were trying to crash it. I managed to stabilize the flight, and I saw the critters falling off the aircraft. I don't know if they fell and died, or if they jumped from my plane to a different one. I have no idea. So, again, let's assume that the Gremlins aren't really real, and the product of some other more explicable phenomena. And let's put aside the observation that they're pretty good examples of a trickster archetype. I'm sure we all know what that is, or we can Google it later. Today, we're interested in what could cause the myth of the Gremlins to come about, and what purpose it served. The most widely held theory is that oxygen can be in short supply at high altitude, and lack of it could easily cause hallucinations. Other writers suggest that the Gremlins served a morale-preserving function, allowing aircrew to blame faults on something other than the maintenance crews, their comrades. I think there might be something else going on here, too. A military plane is a fiendishly complex piece of engineering, with an awful lot of things that can fail in a lot of unpredictably interrelated ways. In other words, hard to understand. So, once you hit a certain level of complexity, why not Gremlins? It seems as plausible as anything else. It's like the dark side of that famous quote about sufficiently advanced technology that I won't waste any time about by repeating here. And I think we can see this effect at work when we think about the stories that people make up to explain technology to themselves. Julie Carpenter, writing about myth-making, says this. She's talking about how groups of people start to create stories when they don't know how something works. The thing I liked about numbers stations wasn't just that they represented this tantalizing powerful mystery of the radio network, but they also drew people together to try to figure out what they were doing and why they were there. And these stories later intersect with the internet, the next generation of network. All kinds of weird things fly off. As people's photos appear, allegedly of site visits to defunct stations, strange YouTube channels and Twitter bots appear, extending on the idea of hiding coded secrets in plain sight. The raw material of the story of the number stations, recordings, logs, theories of who might have been for orchestral and why, get extended on by many imaginative minds. Arthur Frank, writing about how stories work socially, says this. He's talking about how storytellers and people who listen to stories are all working from a kind of shared culture, and storytelling is a sort of process of building that they all do between themselves. He also has this to say. This kind of process of creating a shared story is fundamentally social because you're sharing these fragments between yourselves. It's these social aspects of storytelling that I think gives us the modern meme culture, the shared narrative sources and social functions of a story multiplied hugely by network effects. The story like Slenderman, for example, who first pops up on the Something Awful forums in a Photoshop competition to make super creepy images. His images show a faceless, gangly creature lurking in the woods behind some oblivious kids. However, the story swiftly takes on its own momentum as it comes into contact with many more creative minds, all adding their part to the legend. Playing with fragments of folklore to create an ongoing shared viral narrative, popping up all over the internet in more creepy images and stories implicating the Slenderman in atrocities large and small. And at the same time as being a thoroughly networked fable, Slenderman simultaneously has its roots in our deep folklore. Shira Chess notes that he owes many of his characteristics and some of his behaviors to fiction and folklore characters that preceded him. She makes a connection between Slenderman, the collective mythos of Slenderman, and that of middle European fairy lore, pointing out that with an AE, pointing out that many of Slenderman's characteristics and behaviors, luring humans into traps, distorted or indeterminate body features, child abduction, and a general air of unheimlich, all have precursors in traditional folk tales. Slenderman turned out to be such a powerful story that it even infected old media. While writing this talk, actually while writing it, I spotted these movie posters appearing along my commute, which freaked me out quite a lot. And I'm sure it's a coincidence. So we've seen the way that we think and tell stories about our technologies, actually has close ties to the ways that humans have always told stories about the world and interpreted phenomena. I wanna spend the rest of this talk examining one particular strand of folk archetype, the Arachula machine. And I'm gonna start with a story of Roger Bacon, a 13th century friar and philosopher, and his alchemic bronze head. According to the records, friar Bacon wasn't the first to cast a head from bronze in order to create an oracle. Apparently Pope Sylvester II made one in the 9th century or stole it with the aid of some demons, but that's another story. Several were said to have been built by Renaissance makers in the 12th and 13th centuries, although written accounts don't appear to the 16th century, make of that what you will. These heads were apparently able, after sufficient engineering and or alchemic effort, to answer any question put to them or make predictions about the future. But some, familiarly, could only answer yes or no. It's the accounts of friar Bacon which became the most famous though thanks to the 16th century playwright, Robert Greene. In Greene's play, based apparently on a true story, Bacon and his assistant Miles spend a great deal of time and effort building a bronze head that in the inward parts thereof, there was all things like as in a natural man's head. Finally, they're ready to animate it, which required keeping a continuous watch and also the continual fume of the six hottest symbols, which are plant extracts to you and me, your basic alchemic building blocks. After three weeks of this, Bacon falls asleep, exhausted. While he sleeps, the head finally boots up and says three things. Time is, time walks, time is passed. Miles, still awake, freaks out so hard that he knocks the head over and it shatters. Bacon sleeps through the whole thing. Years later in 1837, Charles Babbage theorized that given enough information, an entity might distinctly foresee and absolutely predict, for any, even the most, most remote period of time, the circumstances and future history of every particle. He was thinking about a process which could model the whole universe and thus predict the future. Alexa, what is the weather today? Right now in Seattle, it's 71 degrees. Where should we go today? Alexa, what is the weather in Bangkok, Thailand? In Bangkok, Thailand, it's 81 degrees. Sound familiar? The idea of a machine which can answer any question has been with us for at least as long as we've been making machines. The original mechanical turk is in here, too. The robot which actually was a machine The idea that an engineer's ingenuity can tell the future is also interlaced with the Faustian myth of being granted great knowledge through esoteric means, but that knowledge usually comes at a price. I think this archetype is alive and well in the ideas framing the discourse around big data and machine learning, and it's not just me. I think that Alchemy's getting a bad rep here. It did, after all, lay the fact that it did, after all, lay the foundations for a lot of modern science, medicine and chemistry. It did, though, have its fair share of rituals that were enacted because the rituals themselves were thought to be powerful, and this is the analogy that Google's researchers are making. They claim that most people doing AI, are following rituals that they've learned without really understanding the underlying science. Input data sets and training parameters are tweaked according to recipes people find on the internet or learn and accumulate by trial and error. I suppose you could call that folk knowledge. There's another problem in AI. You can't open it up and see what's inside. Sure, we understand what's happening inside a neural network when set the conditions that allow it to create itself, and we know how to measure those outputs and tweak the whole system until we get the results that we want. We can't actually look inside the neural net as it's running, or at least it's very... it's infamously difficult to do a good visualization of that. We don't even know why neural nets are so effective, let alone what's going on in there. We're creating black boxes, but as long as we keep doing the rituals, the systems will behave. It never ceases to amaze me that as an industry we're actively involved in a project of hyping off more and more important data processing and decisions to machine learning systems and neural networks, self-driving cars for sure, but also setting insurance premiums, making healthcare decisions, setting bail, predicting someone's chances of offending, running the stock market. And not the heart of all of these systems? Faith. Faith in unknowable black boxes. This is a quote by William Stahl from his book God and the Chip, which is all about how people think about how religion and technology and how people think about both things intersect. It's a wonderful book if you're interested, get it. He goes on to say that there are dangers in treating technology like magic. Scientists and engineers become priests and wizards whose pronouncements cannot be understood by the public and implicitly should not be challenged. So, I could end here on a pithy summary that wraps everything in a neat little bow, but that would seem like cheating. We like neat resolutions, happy endings, elegant reversals, thematic resonance. We like frameworks, and that leaves us susceptible to a kind of narrative exploit. If a story seems to fit a pattern that we know, we'll accept it more easily because we know that story on a fundamental level. Superstar CEOs are a neat fit for the hero's journey. A new shiny thing is a gift from the gods, powerful and capricious. All-knowing AI systems become minor deities capable of doling out judgment on as imperfect mortals. Deployment of these archetypes, doles are creative faculties, makes us more prone to magical thinking, absolves of its responsibility as we get swept up in the narrative. And just as we like our stories to follow orders and structure, we like to think that our technology is a product of science, rigorous, rational thought. But they're also the products of people. People like you and me sitting down at a keyboard or a drafting table or a drawing board and bashing out their opuses. And those people are all embedded in a culture. They grew up on particular stories. We can't help but have the way that we think shaped by those stories. In a way, those stories are what taught us to think. And we repeat those patterns without even knowing it. So I'm not going to end this neatly, but if I am going to leave you with one thought, it's this. Keep alert. Reserve special skepticism for those stories which seem too neat, too powerful, too universally true because they might just be the regurgitation of our shared folklore to the bamboozlement of all involved. Thank you. That's me. That's some links. Thank you. Perfect timing. We have time for a few questions. As I mentioned before, I have this wonderful microphone. So if you have a question, this microphone will fly to you. Raise your hand if you have a question. Please. Anyone? Yes. Have you ever tried inventing your own folk story? Have I ever tried inventing my... Have you ever tried inventing your own story? Well, strange you asked that. On the badges, you should be able to download an app called 23 Million Stories, which generates 23 million short stories. Not all at once. But yeah, I have done quite a lot of work before this talk. I work in generative language, Twitter bots, that kind of thing, generative short stories. I haven't tried to make my own universally true bit of folklore yet, but you know, give me time. Any further questions? Yes, over there. Could you pass the microphone over to the gentleman over there? Just throw it in the general direction and people will hopefully pass it on. Yes, you got the concept. Thank you. Give it a second. Try again. Hi. Hi. Do you have any predictions about what kind of folklore we may see in the future, given current technology? That's a good question. I think... So, there's a big thing in systems design where we taught that making things seem seamless, things seem natural in the UX and UI and the flow and everything. We taught that that's a good thing, right? We should try and hide complexity from... We should try and hide complexity from people, right? We should try and make things appear magical. And I think as systems get more complex and we hide more stuff, then the amount of stuff that people have to make up to try and understand what's going on gets bigger, right? So I guess there's that. There's also the... Yeah, there's kind of... I wouldn't want to hazard a guess at what specifically might happen, but we'll see the popular imagination of how things work from how they actually work. I think that Delta's only going to get bigger as systems get more complex and we hide more of their details. So, you know, don't do that. Any further questions? If so, please raise your hand so we can... Oh, all the way the way. Could you pass the microphone all the way to the... Whoops. Thank you. Do you see value in leaning into the folklore? Oh, yeah, absolutely. Okay. I mean, to contradict exactly what I just said. So my work work, I worked for the BBC in research and development, and I've spent the last couple of years looking at voice interactive devices, so stuff like Alexa and Google Home and whatnot. And the thing about those devices is that they're really dumb, right? They're sold on the promise of being this all-knowing AI, but actually when you get into developing stuff for it, it's not really AI. We're making like decision trees and big branching if-then-else statements. But they are sold on the promise of being an all-knowing AI. And if you're good with your writing, you can totally play on that and pretend that your system knows more than it does. And we've made a few games that do that. Like, we've made a few games where we write AI characters that appear to be smarter than they are and just weird sort of magician stage trick kicks in, where if you're sort of manipulating enough with the way that you use language, you can totally imply that your system's doing more and I also think that as a sort of... There's a sort of... What's the word? I think it's also useful for sort of activism and presenting an alternative vision of how things might be. I went to a thing called Haunted Machines at the Impact Festival last year. There's a guy there called Scott Smith and he was talking about how if you're good enough about telling your sort of alternative visions of the future, if you get really good at telling that story, then that story becomes a weapon against the stories told by, for instance, governments you might not like or companies that you might not like about the future and it becomes this kind of like really powerful totem if you tell that story really well, because people love stories. So it becomes a war of who can tell the best stories, which I'm pretty sure my mate Sharkle will be talking about a bit later. I'm done. So if you could pass the microphone back to the front if you would, to the kind and all the others, please give a big hand of applause to Henry, who has to do this talk again in two years. Thank you.