Entertainment
Imagine to wake up in Rome 2000 years ago, on the Nile in ancient Egypt or in the streets of London during black death in the 1300s – with realistic sites, sounds and daily struggles.
In recent weeks, videos generated by AI showing the views of people who wake up in different historical periods have become viral on Tiktok.
Dan and Hogne are the creators behind two of the accounts – POV Lab and Time Traveler Pov.
Dan, based in the United Kingdom, tells the BBC that he creates these videos as “the idea of seeing the past through a first-person perspective seemed to be a unique way to give life to history”.
Hogne, a 27 -year -old man from Norway, adds that his videos teach people “cool parts of history and help them learn something new”.
Despite the videos offering millions of windows on history, a number of historians have shared concerns about the accuracy of the content and if AI can really resuscitate the past, or we just see a polished and modernized version Designed for commitment?
Allow Tiktok content?
The troubled and misty streets, the coughing of cities and the distant toll of the bell of a plague doctor are all in the most watched video of Hogne, which has accumulated 53 million views.
This has aroused fascination among many, but the historian Dr Amy Boyington describes video on the medieval theme as “amateur” and “evocative and sensational” rather than historically exact.
“It looks like something of a video game because it shows a world which is supposed to look real but which is actually false.”
It underlines inaccuracies such as the representation of houses with large glass windows and a railway which crosses the city which would not have existed in the 1300s.
The historian and archaeologist, Dr. Hannah Platts, has also noticed important inaccuracies in a video illustrating the eruption of Vesuvius in Pompeii.
“Due to the account of the eyewitnesses of Pliny The Younger of the eruption, we know that it did not start with lava that spits everywhere so as not to use this richness of historical information at our disposal inexpensive and lazy. “
Allow Tiktok content?
She also spots smaller errors – rod wine glasses and pepper mills did not exist and people would be lying while eating unaccompanied on dining chairs.
“The jet of bread in the video is a modern bread and since we have charred breads from the moment it is a shame that the person who was doing the video does not do any research and does not understand that. “”
Dan, who created Pompei’s video, says that he recognizes a lot of details in his videos is historically inaccurate.
“The content generated by AI is not perfect, and even if I seek precision, these videos consist more of evoking the feeling of a period of time rather than being a 100%factual recreation.
“They are more like artistic interpretations rather than strict documentaries.”
‘Manipulate the story’
But, Dr Boyington is concerned about the impact that these artistic interpretations can have on the rewriting of history.
“It can be quite dangerous because people could handle the story – for example, someone could create a video generated by AI that withdraws holocaust traders.”
While most people will know that the content of the video is not real, the concern is for “young people who will find out about a historical period for the first time through these videos”.
Dan rejects these statements and says that his videos “are not supposed to be considered a pure historical fact”.
“I encourage viewers to seek the history themselves if they are interested. I see these videos as a way to arouse the curiosity of the past rather than replacing a real historical education.”
Hogne says that he feels “a responsibility” in the creation of these videos and focused on making the videos as precise as possible “, especially now that so many millions of people are watching them”.
Dr. Platts is concerned about the disinformation that spreads without control, noting that certain viewers in the comments did not seem to know that the videos generated by AI were not based on historical facts.
“We see many students who now use AI, and what is problematic is if they see something like that and then it is echoed as if it was a fact.”
Dan says he ensures that all his videos are labeled as created by AI, while Hogne says that disinformation existed long before AI and “people must think in a critical way of everything they look at “.
‘Can have immense advantages’
All the historians to which the BBC spoke is suitable that there are merits to the videos of Dan and Hogne.
Dr. Boyington says that they can act as a “gateway in history and can inspire someone to do their own research”, while an Egyptology teacher, Elizabeth Frood, says that “if it is done with Precision and reliably, then this is of a huge advantage for the public as a public as the public as a public as the public as a public like the public like the public like the public like the public as the public like the public like the public like the public by As long as the public as the public as the public as the public as the public as the public as a public as the public as a public as the public as a public as the public as a huge advantage for the public As a public as the public as a public like the public as the public as an immense advantage for the public as a public as the public as a public as the public as a huge advantage for the public as a public as the public as a huge advantage for the public as a public to the public for the public for the public. It arouses the interest and the conscience of history “.
Barbara Keys, professor of American history at the University of Durham, took a look at a Video generated by Ai-Ai Someone working at the Chernobyl nuclear power plant on April 6, 1986, the day the reactor exploded.
Although she sees the potential advantages of these videos, she calls the clip a “black box” because there is no source information or transparency on the information that has been transmitted to the AI program.
She says that a worrying inaccuracy in the video is that the retractor is based on images after disaster rather than before.
“It makes people think that Soviet technology was really bad when it was actually very sophisticated.”
She thinks that the viral nature of these videos concerns their historical content less and more on the fascination of the public for the AI itself.
“There is nothing very interesting in the video, and that gives you no information on the accident or what happened afterwards, so the interest must be capable of what AI can do. “”
Professor Frood fears that the videos have not been created with precision or reliably and stresses that historical reconstructions should be based on in -depth research and verifiable sources.
“We have to criticize this video because we know nothing about its source or what information has been fed at AI for this video to be created.”
She specifically comments on a video to be a child in Egypt in 1250 BC.
Allow Tiktok content?
The first thing that struck her was how videos tend to “homogene the old complex worlds”.
“Egypt extends over thousands of years, and people may not realize that it is just snapshot,” she explains.
Overall, she says that the video seemed to be “badly studied” and underlined a case where a teacher in a school scene reads hieroglyphs back when he should be read from right to left.
Hogne, who created the video, says that he recognizes that there are historical inaccuracies in his videos and it is because “AI can make many mistakes but in the future, the tools will improve “.
“I try to spot them where I can but I am not an expert in history, so I do not always see the error, especially when it is something small,” he adds.
Asked about the lack of transparency in source equipment, Hogne says in the future that he could consider adding links at the origin of her information.
“It would be cool to make it completely precise using the right sources, but it’s just me by my own videos and it takes eight hours to do each.”
How do you create these videos?
Hogne relies on the GPT cat to seek a period of time and requests her information on people and places would have felt. He then generates an initial image of someone who looks at a landscape and from there “it’s a lot of back and forth to do things well”.
Dan says that he takes four hours to make each video and that he uses “different AI tools to generate high quality images, animate them and create a realistic sound”.
He watches videos and reads documents to learn over a period before making the video to try to give him “a certain historical clarification”.
“You have to give all the details to AI, from” traditional Italian dresses “to” paved floors “, otherwise it will be creative and will do random things you don’t want.”
