Artificial intelligence is doing some pretty remarkable things these days, but there’s something else happening that perhaps we haven’t given enough thought to. What happens when AI starts interfering with our memories? What happens when we begin remembering things that never actually happened?
An interesting article published by The Washington Post explores this very question, and some of the findings are fascinating, if not a little disturbing.
Researchers at MIT have been studying how artificial intelligence can influence human memory, and what they’ve discovered is that people can actually develop false memories after being exposed to AI-generated or altered images. In one experiment, researchers showed participants ordinary photographs and later presented some of them with AI-altered versions. Those who viewed the altered photographs were more likely to remember details that never existed, and perhaps more concerning, they were increasingly confident that their memories were accurate.
Other studies have found similar results. People questioned by AI chatbots after watching a simulated crime developed significantly more false memories. In another experiment, nearly half of participants falsely remembered movie remakes that never existed, including a version of The Matrix starring Will Smith. Researchers are also discovering that people who rely on AI to help them write or generate ideas sometimes have difficulty remembering which thoughts were originally their own.
Of course, none of this means AI is automatically destroying our memories, but it does raise some interesting questions about the future of human cognition and our ability to distinguish between what actually happened and what we simply believe happened.
Recently, there was this “Tiptoe in Your Jordans” meme that caught on, featuring a strange little troll that looked like something out of an old movie or television show. A lot of people felt that they had seen this little troll before, but they hadn’t.
People got aggressively angry arguing that they’d seen him before. Some were absolutely convinced that this character came from some obscure movie from the 1970s or 1980s. But instead, it was just an AI-generated meme that became incredibly popular and has now created a lot more confusion.
Additionally, there’s something else going on on TikTok. A username by the name of Harold Puddington has a humorous body shape and long neck, and the videos are so good that people are constantly asking if it is AI or not. To the point where AI is now creating a new reality of people that don’t exist, or at least making it increasingly difficult for others to determine whether they do.
A part of the issue, perhaps, is that AI learns from things that have already been there. So, in theory, if it draws from the right things, or the right visual memories of our collective past, and creates something new, people may have a vague association with what they see despite the fact that the particular thing they’re looking at never existed.
Think about that for a moment.
AI can create something that looks like a character from a forgotten children’s television show, an old movie, or even a commercial from 40 years ago. It can have the same grainy appearance, the same awkward movements, the same strange facial expressions and the same atmosphere that makes something seem familiar. And because our brains recognize certain patterns, we may associate that new creation with something we actually remember, even though the creation itself is entirely new.
This complicates things for people who are believers in the Mandela effect, and it certainly complicates memories and false memories and really muddies the waters of everything.
For years, people have argued over the Mandela effect, whether it was the spelling of the Berenstain Bears, the Monopoly man’s monocle, or countless other things that people collectively remember differently from how they actually existed. Now imagine introducing artificial intelligence into that equation, where entirely new characters, people, events and images can be manufactured and circulated across social media until millions of people have seen them.
It’s interesting because our brains are meant to see things and visualize reality and understand and remember. No one’s memory is exactly perfect, but AI is now scrambling the normalcy of all of this.
And there’s another side to this that perhaps is just as concerning. We’re now getting to the point where people are questioning things that actually are real because they look like AI. A real person with an unusual appearance can suddenly become the subject of an online debate over whether they exist. Meanwhile, a completely fictional character can be accepted as real, or even remembered as something from someone’s childhood.
Social media only makes this more complicated. Someone sees something, insists they remember it, and suddenly hundreds or thousands of other people begin agreeing. Some may genuinely feel that they remember it, others may simply be going along with the crowd, and before long an entirely new collective memory begins to develop around something that never happened.
And perhaps that’s the most fascinating part of all of this. AI isn’t simply creating false information anymore. It may be creating an entirely new sense of familiarity, a world where things that never existed somehow feel like things we’ve known our entire lives.
We’re already living in a world where photographs can be fabricated, videos can be manufactured, voices can be duplicated and people can be created out of absolutely nothing. Now we’re beginning to understand that these creations may also have an influence on how we remember our own past.
This is not the end of anything or even the middle, this is just the beginning. AI is going to increasingly get better to the point where maybe one day we’ll actually see Harold Puddington in real life tiptoeing in his Jordans with a tiny troll that will make gold.
Unless we saw that before. 🙂
