When AI Looks Almost Human, “Almost” Becomes the Problem. Have you ever watched an AI-generated person who looked realistic, but something still felt wrong? The face looked human. The voice sounded human. The movements were close to human. But your brain still whispered: That is not a real person. That uncomfortable reaction is known as the uncanny valley. 🌀 What Is the Uncanny Valley? As robots, avatars, and digital humans become more realistic, people generally respond to them more positively. But there is a point where something looks almost human without being convincing enough. At that point, our comfort level can suddenly drop. Small imperfections become strangely noticeable: - Eyes that do not move naturally - Facial expressions that arrive a fraction too late - Lips that do not perfectly match the words - Skin that looks too smooth - A voice that lacks natural emotion - Body movements that feel slightly mechanical A cartoon character does not bother us because we know it is not human. We are not expecting perfect realism. An AI avatar that looks 98 percent human creates a different expectation. Our brains become extremely sensitive to the missing 2 percent. ⚠️ Why This Matters for AI: AI-generated video, voices, virtual assistants, and digital employees are improving quickly. The goal, however, should not always be to make AI indistinguishable from a human. Sometimes a clearly artificial character feels more trustworthy than a digital person pretending to be real. This creates an important design question: - Should an AI experience try to look completely human, or should it be honest about being AI? There is no universal answer. It depends on the audience, purpose, and level of trust required. The uncanny valley reminds us that technical realism and human comfort are not the same thing. As AI becomes more humanlike, the smallest details may determine whether people feel connected, curious, uncomfortable, or deceived. Share in the comments:

I really enjoyed your conversation with Herman. It takes courage, especially with how quickly AI is changing. I think a lot of us feel out of sync at times. But then there are moments when I hear someone else speak or read something, and it reminds me how much I actually do know, even if I don’t always realize it. This format has been great because I can jump in whenever I feel like it—whether to learn something new or revisit what I already know. I hope that makes sense..

I decided to have some fun and put down Claude Cowork and Code for the weekend and deep dive into ChatGPT image generation. I needed a catalyst, so I entered a Product Image Challenge, here. The challenge gives you a different real-world product each day. Your job is to turn it into an eye-catching product image or advertisement. For my entries, I’m using ChatGPT with Sol Light to develop the concepts and create the images. One thing I’ve found interesting: the lower models with heavy thinking, or higher models with lower thinking can work really well for creative projects where you’re looking for a general representation of an idea. You don’t necessarily need the biggest model to start creating. The harder part comes when you want consistency of characters, like the series I am creating, called “Milo.” DAY 1: Coca-Cola 🥤 This was more about creating an interesting scene and finding a visual idea that worked with the product source image. This was where I thought of Milo. DAY 2: Heinz Ketchup 🍅 For Heinz, I created a standard entry, and then went up and beyond and started experimenting with recurring characters and visual storytelling. And that introduced a completely different challenge. DAY 3: For GoPro, I created a continuation of the Milo series working with perspective to create an emotional and memorable futuristic scene. 👉 How do you keep the same characters looking like the same characters from one scene to the next? Same little girl. Same humanoid character. Different location. Different pose. Different action. That sounds simple. It isn’t. 😂 You quickly discover that prompting an AI to create a great-looking character once is very different from trying to direct that same character through a series of scenes. That’s where reference images, detailed descriptions, iterative prompting, and a little patience start becoming important. I’ve attached some of my creations from the first two days, and the source images.

In Part 1, we compared Chat to Dory and (Co)work to Spock. Chat feels conversational and creative. (Co)work feels structured and logical. But this is not just a personality difference. It is also a feature difference. 🐠 Chat is primarily where you talk with ChatGPT or Claude. You ask a question. Explore an idea. Refine responses. Chat can search the web, use enabled connectors, analyze uploaded files, and create content. However, you are driving the process one response at a time. 🖖 (Co)work is designed to complete work. You can point it toward a working folder, give it access to relevant tools, and assign a broader task. For example, (Co)work can: - Read multiple files from a connected folder - Create new documents and reports - Update or reorganize existing files - Rename, sort, and categorize documents - Gather information from connected applications - Use a browser or desktop application when required - Complete a task involving several steps - Run recurring scheduled tasks The practical difference is this: Chat may help you write the plan; where (Co)work can open the working files, follow the plan, and produce the deliverable. Imagine that I am working on a book. In Chat, I might say: “Help me develop the central argument for this chapter.” We can brainstorm, challenge ideas, test language, and find the story. In (Co)work, I might say: “Review the chapters in this folder, identify repeated sections, update the chapter outline, and create a revised draft.” One helps me think through the work. The other can work across the actual materials. They are designed for different purposes, that is why they feel different. Chat is optimized around the conversation - scratch pad of ideas. (Co)work is optimized around the assignment - execution of ideas. Of course, the lines continue to overlap. For example, Chat can use tools, and (Co)work can brainstorm. But their centers of gravity are different.

Your feedback is helping shape the future of AI Bits & Pieces, and I appreciate everyone who took the time to vote. Top Three Topics, by votes: 🥇 Introduction to ChatGPT Work / Claude Cowork — 5 votes 🥈 What Is an AI OS (Second Brain)? — 3 votes 🥉 How I Used Obsidian to Create an AI OS — 3 votes We’ll begin creating content around all three topics. 🎁 Giveaway Winner! Congratulations to @Garry Cole, who was randomly selected by AI from all eligible participants! Garry will receive a $25 Amazon eGift Card. ❤️ Winning Charity The winning charity is Winning Futures (@Kristina Marshall)! As promised, AI Bits & Pieces will make a $25 donation to Winning Futures in honor of this giveaway. Thank you again for being part of this community. Congratulations to Garry Cole and Winning Futures! 🎉