“I keep having these memories […] but I can’t tell if its real.”
That’s a quote from the character Carol Danvers – aka Captain Marvel – played by Brie Larson in the 2019 comic book movie of the same name.
But what if you were told the film Captain Marvel was remade in 2020, starring Charlize Theron in the leading role?
Psychologists from University College Cork in Ireland were investigating growing concerns about the potential for deepfake technology to spread misinformation and distort memories.
They provided 436 participants with deepfake remakes of popular films like Captain Marvel, The Matrix and Indiana Jones.
The researchers found 73% believed the fictitious Captain Marvel remake (starring Theron) was real, and 41% of those actually said the pretend film was better than the original.
Deepfakes use artificial intelligence (AI) to create highly convincing, but false images, audio or video.
In the study, participants – who thought they were being surveyed about recent movie remakes – were presented with four real and two deepfake movie remakes in random order. Three movies were presented as deepfake or real movie clips, and three were presented as a short text description.
Fictitious remakes included The Matrix (starring Will Smith), The Shining (starring Brad Pitt and Angelina Jolie), Indiana Jones (featuring Chris Pratt) and Captain Marvel (starring Charlize Theron).
Real movies included Charlie and the Chocolate Factory, Total Recall, Carrie, and Tomb Raider.
For each of the six movies, participants were asked whether they had seen or heard of the original film, and/or the remake.
The study shows participants readily formed false memories of fictitious movies, with around half (49%) believing each remake was real.
The format – whether the fake film was described via video or text – did not have an impact on false memories. Nor did this affect memories of genuine remakes.
Later, participants were told the real purpose of the study – to examine false memories from movies – and it was revealed some of the movies were doctored using deepfake technology.
Participants were then asked whether all the movies they had seen were real, rating them from 0 (definitely not real) to 100 (definitely real).
Once participants knew some of the films were misleading, their judgement of whether films were real or fake improved, with real films scoring higher for truthfulness compared to deepfake versions.
When asked their views about using deepfake technology to replace actors and actresses in films, around a third (31%) of participants were happy to see wider use of the technology, compared to more than half who were not (55%).
Popular reasons for doing so included “If the replacement actor better fitted my perception of the character” and “If I found the original actor annoying”.
The study is published in PLOS ONE.