This text belongs to ‘Artificial’, the newsletter on AI that Delia Rodríguez sends out every Friday.
The video starts strong. “My body disintegrated at the bottom of the ocean and all to please my father,” says the hyper-realistic image of a young man in an orange sweatshirt, looking at the camera and with an underwater scene in the background, who continues to narrate: “My name is Suleman Dawood and this is my tragic story.
For more than two minutes the video, clearly generated with artificial intelligence and so striking that it is difficult to take your eyes off it, continues to imagine in first person how the 19-year-old who died along with his billionaire father in the submersible that died would have felt. he wanted to visit the remains of the Titanic, his thoughts, his last minutes and his feelings towards the man who gave him the trip without knowing that they would not return from it.
The video has more than 67 million views on TikTok, a spectacular number for this platform. The account that published it less than two weeks ago, Human Stories, which promises “real stories told by its protagonists”, has more than a million followers. There are dozens of very similar accounts. @yestaesmihistoria also exceeds a million. Others like @historiasvividas, @estaesmihistoria_ia or @memoriasinmortales exceed half a million.
The viral about Suleman Dawood is the most successful I have found of a new genre that is triumphing on TikTok: “AI storytelling”, the narrative built with artificial intelligence, which adds new tools with some old audiovisual tricks to get images shocking synthetic animated. His hashtag on TikTok has more than 13,000 million views.
Both the video about Dawood and the rest of the videos published by this type of account share certain characteristics. Its protagonists are victims or villains of some of the best-known events in history, although the horror genre and the autobiographical experience of great characters are also used. The narrators explain their story in first person and looking at the camera in what looks like an animated photo, probably built with MidJourney or some tool that provides lip movement and voice, like HeyGen or Studio D-ID. The script also appears to have been created with artificial intelligence, says Victoriano Izquierdo, co-founder of Graphex and a photographer who has created AI images for The New York Times.
From the girl Asunta Basterra to Cleopatra, from the woman who killed all her children as she was having them to Josef Fritzl, the more or less real stories are addictive. The commentators don’t seem to care as much about the format, which they seem to accept as something as natural as a cartoon, as the gruesome details of the narrative itself. “They should give you an award”, some people answer in the comments to someone who dares to break the magic of the story to call attention to the fact that the woman who speaks does not look like the real victim of a crime or that the Peruvian girl who became a mother at the age of five was a brunette and not a blonde. The combination of TikTok’s algorithm and generative artificial intelligence is a bomb. If the narration captures your interest for a few seconds (how to avoid it?), the platform will show you similar questions over and over again.
The new genre is very recent. It was only at the end of May that some mentions began to appear in the press, with an article in Rolling Stone magazine warning of the rise of artificial true crime staged by children. However, this Anglo-Saxon scene (Kotaku was surprised by a video with 100,000 views) pales in comparison to the success of the much more viral videos in Spanish. The phenomenon is not strange: despite having an audience theoretically more limited by language, certain narratives related to storytelling have always been more shared on Facebook among Hispanics. We seem to really like a good story, no matter its format or our country of origin. Similar hits may exist in English, but we haven’t found any videos or accounts with a comparable reach to the ones we’ve cited.
For Janira Planes, an expert in digital communication and creator on TikTok, these videos remind her of another hypnotic genre on the platform: videos where great stories from the Reddit forum are superimposed on images, for example, of Minecraft. The way in which the authors can monetize the content is not clear, but some promise to explain how to make money by creating viral videos with AI tools.
I’ve been researching these videos for several days. I have not been able to contact the authors, but they could be anyone: I myself tried the tools that we have mentioned and it took me about twenty minutes to generate a quite nice Salvador Dalí explaining how he took his ocelot to parties. I’m not very original, nor are the creators who have started to play to create series like “if Harry Potter happened in Italy”.
I do find narrations of events in the first person more interesting, no matter how atrocious, disturbing or sensationalist they may be: we have in our hands very sophisticated tools and our imagination returns to the stories around the stake and the crime, with its warnings and its morals, to basic storytelling techniques. We are not able to resist a first sentence like “my body disintegrated at the bottom of the ocean and all to please my father.” Especially when we witnessed the event and it is about our own shared mythology, our stories of the 21st century.
What else has happened this week
– Nature has published two Chinese artificial intelligence models that predict the weather with an accuracy similar to that currently achieved by meteorologists, but much faster.
– Japan has gone ahead by developing a guide on how tools like ChatGPT should be used in the classroom. For example, they should not be accepted during an exam, but they should be useful in language classes. It is recommended that teachers and students talk about its uses, its risks and its ethics. British universities have also developed standards of behaviour.
– It may be just an excuse, but Elon Musk limited the use of Twitter last weekend with the argument of preventing access to his network by the big data extraction models. Meanwhile, Meta has launched its competitor Threads, although it cannot be accessed in Spain.
– Junts per Catalunya has created several campaign videos using AI deepfakes that are more disturbing than anything else.
– The war of the chips continues. China responds to its limitations by restricting the export of basic metals such as germanium or gallium for the manufacture of semiconductors, cars or defense equipment. And it appears that the US is thinking of requiring prior authorization from the big platforms before offering their AI chip-powered cloud services to Chinese customers.
– The Wimbledon organization is using AI to comment on match videos.
– From curing all diseases to fighting climate change or making faster progress in all disciplines, The Guardian asks five experts how AI can help us. In English.
– Important: two writers have denounced OpenAI, creator of ChatGPT, because the summaries they make of their books are accurate, so they understand that they have been used to feed their language model without their permission. In English.
– Microsoft and LinkedIn launch a free introductory course on generative artificial intelligence, in English.
– Creator Nick StPierre sums up Midjourney’s astonishing improvement in just one year in one tweet and two images.
– Interesting: eight companies are managing to attract scarce talent in AI. They are Open AI, Anthropic, Hugging Face, Midjourney, Stability AI, Character AI, Jasper, and Tome. In English.
– The actions of pressure on the future AI law continue. Some 150 CEOs and presidents of large European companies have written an open letter to the European institutions asking that it be diluted so as not to endanger European innovation and technological sovereignty. They are Airbus, Siemens, Renault, Heineken or Deutsche Telekom…
– Does anyone else remember that old machine in the shape of the Roman Bocca della Verità that read the future for a few coins? He’s back, and now with AI and using astrology! (In English, in The New York Times)
Anxiety level this week: medium, I am very entertained using corporate video animation tools with great characters from the story (they don’t work with pets yet).
