SYSTEM NOTICE

Auto translation by AI. Be sure, accuracy, nuances and authorial intent may not be fully reflected.
見出し画像

3 Reasons Why a Psychologist Was Chilled by AI Detection with Only 70% Accuracy

Have you ever hesitated for a moment while watching a video on social media, wondering, "Is this real?" AI-generated videos have become surprisingly realistic over the last few years. The shape of the fingers and the subtle shifts in facial expressions are now almost indistinguishable to the untrained eye. However, the truly frightening part is not the videos themselves, but the idea that as we continue to watch them, our innate "sense of discernment" is quietly slipping away. As someone who has dealt with the workings of the human heart in a medical setting, this is by no means a matter that concerns only others. Today, I would like to think about this phenomenon with you from a psychological perspective.

Footage that has become something else in just two years


In the spring of 2023, an AI-generated video of "Will Smith eating spaghetti" became a hot topic on social media. The face was distorted, the mouth movements were unnatural, and it was at a level where anyone could tell at a glance that it was an AI video.

However, it is said that in just about two years since then, it has evolved to a quality that even professional eyes cannot distinguish from the real thing. This is what Daisuke Furuta, Editor-in-Chief of the Japan Fact-check Center, said in an interview with Diamond Online. In the past, there were clear "AI-like" traits, such as unnatural fingertips or eyes, and one could tell immediately by looking at those. But with high-performance video generation AIs being released one after another, the current situation is that almost no such clues remain.

When I heard this, I didn't think of it as just a story about "technology advancing." Instead, my interest turned to the possibility that our own way of perceiving things is quietly beginning to change.

There are far more people who spread misinformation than those who create it


Another thing that left an impression during this interview was the categorization of those who spread fake or misinformation. It is said that spreaders can be broadly divided into three types: those who spread it intentionally for profit, those who believe they are the ones who are right and spread it, and those who spread it seeking attention or approval.

However, what is even more troublesome is that the number of people who believe this is "correct information" and spread it is overwhelmingly larger than the original spreaders themselves. Because there is no malice, it is harder for the brakes to work. I felt that here lies a structure where the damage chains on indefinitely.

From a psychological perspective, this is a very natural phenomenon. If the human brain were to examine every piece of information it sees from scratch, it would never be able to keep up with the processing. Accepting things that are "familiar" without suspicion is a kind of energy-saving mechanism for living efficiently with limited energy. This is not because one has a weak will or is careless. It is a very natural function inherent in everyone's brain.

The scariest thing is not AI, but "habituation"


What made me think even more was the fact that even if you use AI to determine the authenticity of a video, the accuracy is currently only around 70 to 80 percent. No one would willingly eat food that has a 20 to 30 percent chance of being wrong. That is how much of a work-in-progress this technology still is.

However, what Mr. Furuta was most concerned about was not the detection accuracy itself, but the changes that occur in people as they continue to watch AI videos. Many people gradually give up on questioning whether something is real or fake. As a result, the very sense of distinguishing between real and fake gradually becomes dull.

Originally, footage taken at the scene of a disaster or war is inherently rough and imperfect, with shaky screens or things obscured by smoke. On the other hand, videos created by AI are polished and dramatic, like a scene from a movie. Ironically, we tend to feel more reality in that "crafted naturalness." From the perspective of sensory habituation, this is a very convincing story. When we are repeatedly exposed to the same stimulus, our nerves weaken their response to that stimulus. Just as our eyes get used to bright light, as we continue to be exposed to "shocking and dramatic footage," our brains learn that that intensity is the "norm."

Three things I am conscious of to protect my senses


I do not want to make what I have said so far a "story to be afraid of." I believe the important thing is to have small habits to protect your senses once you know how this mechanism works. I will share three things I am usually conscious of.

The first is to stop for a moment, especially the moment your emotions are stirred. Surprise, anger, sadness. When strong emotions are stirred, people tend to hit "like" or share before verifying. Just stop for a few seconds before taking action. That alone makes a big difference.

The second is not to limit your information sources to just one. When deciding on a place to stay on a trip, few people would decide immediately based on reviews alone. You look at the official website and you look at the ratings on map apps. The same goes for information; having several relatively reliable sources will support your judgment.

The third is to have a perspective that consciously distinguishes that "looking real" and "being real" are different things. The more polished the footage or the more dramatic the development, the more you should stop and be suspicious. I believe this is not about becoming a cold person, but rather a gentle habit to protect your own senses.

Summary


I will organize what I wanted to convey today once again.

In the last few years, generative AI video has evolved to the point where even professionals find it difficult to distinguish from reality. Misinformation is spread far more often by people doing so without malice than by the original posters. And the greatest danger is not the technology itself, but the fact that we are gradually letting go of our 'sense of discernment'.

This is not because your will is weak or your attention is lacking. It is the result of a perfectly natural mechanism present in everyone's brain. That is why there is absolutely no need to blame yourself. Simply knowing this can change how you pause and reflect.

Take a breath the moment your emotions are stirred. Do not rely on a single source for information. Be a little skeptical of videos that look too polished. I believe that by keeping these three things in mind, you will surely be able to protect your 'sense of discernment' moving forward.

Thank you for staying with me until the end, despite the length of this post. I hope to continue delivering information that is, in a good way, 'a little bit geeky'.

#GenerativeAI #InformationLiteracy #Psychology #MediaLiteracy #CertifiedPublicPsychologist


いいなと思ったら応援しよう!

AIでエビデンスを探る_公認心理師 よろしければ応援お願いします。いただいたチップは、執筆のための書籍購入や、AIのために使わせていただきます。