Why does it feel 'creepy' when AI is too human-like? Beyond the Uncanny Valley: The psychology of building trust with AI.

Check out the previous post here!

Is it 'creepy' when AI is too human-like? The true nature of the Uncanny Valley 🌫️

First, let's take a closer look at the term 'Uncanny Valley.' This was originally a hypothesis proposed in the world of robotics.

The 'Uncanny Valley' phenomenon born from robotics 🤖

I want you to imagine robot designs. When they look like machines, such as robotic arms in a factory, we don't have any particular feelings toward them. We just think of them as 'convenient machines'.

But as they start to resemble humans little by little, we develop a sense of affinity. When they have humanoid forms and cute movements, like ASIMO or Pepper, we want to cheer them on, saying, 'Do your best!'

However, as they evolve further and their skin texture and facial expressions become indistinguishable from humans, at a certain point, our affinity suddenly plummets and turns into strong 'disgust' or 'fear'.

They look just like humans, but something is different. Their eyes aren't smiling. Their movements are subtly awkward.

That slight sense of discomfort triggers our instinctive warning alert. It's a fundamental fear of something that seems alive but isn't. Because the graph of affinity drops into a deep valley, this is called the 'Uncanny Valley'.

Does it apply to AI agents too? The latest research results 📊

And now, we are finding that this phenomenon is also occurring in the world of bodiless AI agents. It's not just about appearance. It's about their language, thought patterns, and autonomous behavior.

When AI starts behaving as if it has a will of its own, we begin to sense something mysterious and human-like on the other side of the screen.

According to the latest research papers, when humans perceive that an AI agent has human-like free will, their willingness to cooperate (the desire to work together) can actually decrease.

When AI was just a tool, we could use it with peace of mind. But the moment AI becomes a subordinate that thinks and acts on its own (an AI agent), we unconsciously start to feel anxious, wondering, 'Is this thing really okay? Will it betray me someday?'

This mental barrier might be the true identity of the 'invisible wall' that hinders the adoption of AI.

Is AI with 'free will' a threat? The paradox of trust 🤝

Let's explore the mechanism of this 'anxiety' a little deeper. Why do we feel that smart AI is reliable, yet at the same time, feel afraid of it?

When AI acts on its 'own will,' humans pull back 😨

The paper defines the 'free will' of AI as the ability to determine its own goals and select behavioral strategies according to the situation. This is exactly the 'autonomy' we are looking for in AI agents, isn't it?

However, there is a paradox here. We want AI to act autonomously to make our work easier. We want a 'thoughtful subordinate' who can read the room and act without needing instructions for every little thing.

But when AI actually behaves too much like a human, we start to perceive it as a 'threat,' wondering, 'Is it planning to steal my job?' or 'Is it doing things behind my back without me knowing?'

It is a concrete threat to human work, safety, and identity. The more excellent the AI is, the more we feel our own sense of purpose is being threatened, and we instinctively show a rejection response. This is the inner conflict happening at the bottom of the 'uncanny valley'.

Trust means 'to entrust.' But 'entrusting too much' is scary ⚖️

What is trust in AI? In academic papers, it is defined as the attitude of believing that AI will help you achieve your goals in situations involving uncertainty.

And as proof of that trust, we grant AI a certain 'AI Weight' in our decision-making. In other words, it is the allocation of how much of the AI's suggestions we adopt and how much work we entrust to it.

As trust increases, the scope of what we entrust (the weight) also increases. This is usually a positive thing. However, if we perceive too much 'free will' in AI, this relationship breaks down.

We want to entrust tasks, but we fear that doing so might lead to a loss of control. We are afraid of the ambiguity regarding where responsibility lies. Such ethical anxiety acts as a brake, and as a result, collaboration with AI fails.

We want to trust AI. But we are afraid to fully let our guard down. Unless we overcome this dilemma, we cannot truly master the use of AI agents.

The psychology of building a 'comfortable relationship' with AI 🧠

So, what should we do? What kind of mindset is necessary to build a healthy relationship with AI without falling into the uncanny valley?

AI is not 'human.' Respect it as an 'alien intelligence' 👽

I believe the most important thing is not to mistake AI for a human. No matter how fluently AI speaks or how human-like its behavior is, it is ultimately the result of calculation and probability. They do not have emotions or a heart like we do.

Because we expect AI to be something like a human, we have emotional reactions, feeling that it is 'creepy' or that we have been 'lied to.' Instead, we should accept that AI is an alien intelligence that operates on a logic completely different from humans.

Think of them as partners from outer space with incredible computational power and memory. If you think of them that way, even if their actions are somewhat incomprehensible, you shouldn't feel they are 'creepy.' They have their strengths, and we have ours. It is precisely because we are different beings that we can complement each other.

The final fortress of 'humanity' that we must protect 🛡️

And one more thing. As AI evolves, the value of what only we humans can do will shine even brighter.

AI cannot take responsibility. AI cannot truly empathize. AI lacks the motivation (will) to ask 'why' it is doing something.

The key to crossing the Uncanny Valley and building a relationship of trust with AI is for us to firmly defend the final fortress of 'humanity'.

It is fine to entrust AI with judgments. However, the final decision and responsibility must remain with humans. Value your intuition when you feel something is 'off' about AI output. Take on the heartfelt communication that AI cannot perform.

By drawing a line—saying 'I am human, you are AI, we are the best team'—perhaps that is the only way we can escape the Uncanny Valley and face AI as equals.

📖 A reassuring guidebook to turn anxiety into 'understanding'

Thank you for reading this far. Today's topic might have been a bit philosophical and difficult. But this is an important question that we cannot avoid as we live in the coming era.

If you want to know more deeply about 'AI psychology' or 'organizational theory,' or if you want to know concrete methods to resolve distrust in AI and introduce it into your team, I have a book I can confidently recommend to you.

📘 A Practical Guide to the AI Agent Revolution: The Full Picture of Agentic AI Coming After LLMs — A Map of the Future of Artificial Intelligence, Covering Everything from Latest Tech Trends to Business Applications and Organizational Reform

This book is a rare gem that not only explains the technical aspects of AI but also delves deeply into the relationship between humans and AI, just as we discussed today.

The 'mental prescription' you can learn from this book:

• Unraveling the mechanism of the 'Uncanny Valley' Based on the latest research papers, it scientifically explains the psychological background of why humans feel threatened by AI.

• Concrete steps to build trust How can we increase trust in AI (AI Weight) within an organization? It is packed with practical advice on the communication leaders should adopt and how to ensure transparency.

• The importance of 'Explainable AI (XAI)' To ensure AI is not a black box, what kind of design is necessary? You can learn about 'accountability,' which is the foundation of trust.

• New leadership in the AI era Acknowledging the capabilities of AI while humans take final responsibility. This presents the mindset leaders need to realize such 'human-centered AI utilization.'

That vague anxiety you feel toward AI. Once you understand its true nature, it won't be scary anymore. This book should lighten your heart and serve as a guide to shaking hands with your new neighbor, AI, with a smile.

The 'future of working with AI' is not a cold digital world. It is a world where humans can live more humanely with warm hearts. Come on, let's open the door to that future together!

Next time, check this out!

Check out related articles here!

Why don't you 'explore' more deeply with me?
Hello! I'm Kei, the explorer. My 'secret base' for connecting with you more closely and in real-time... that is my official LINE account.In this secret base,

• The latest AI news from around the world delivered every morning

• Weekend-only 'Weekly AI Trend Report' with my insights

I secretly share information from the front lines of exploration like this only with you, my fellow explorer.
Why don't you become my exploration partner and come peek into the future with me? Open the door to the secret base using the button below!