The AI ad that proved the point.

The commercial opens on a boy in a bedroom. He looks about eight, maybe nine. The lighting is warm. The framing is cinematic. And something is immediately, undeniably wrong. His face shifts between frames. His proportions aren't quite stable. His eyes have the glassy sincerity of a stock photo brought to life by someone who's never met a child. This is the Toys 'R' Us commercial made with OpenAI's Sora, and it is the most useful piece of advertising criticism I've seen in years. Not because it's good. Because it's instructive.

The ad debuted with considerable fanfare. A legacy brand, reborn from bankruptcy, using cutting-edge generative AI to tell its origin story. The press coverage positioned it as a glimpse of the future. The audience response was something else entirely. People didn't just dislike the ad. They were unsettled by it. The uncanny valley wasn't a theoretical concept here. It was a viewing experience. And the collective recoil told us more about what audiences actually value than any focus group ever could.

What the audience saw

The technical problems were obvious. Characters changed appearance between shots. The founder, Charles Lazarus, looked like three different people across the sixty-second runtime. Hands had the wrong number of fingers in some frames. Backgrounds warped and shifted. Toys appeared and disappeared without logic. These are known limitations of current generative video tools, and they'll likely improve. But the deeper problem wasn't technical. It was emotional.

The ad was telling a story about childhood wonder. About a man who loved toys so much that he built an empire around them. That story requires warmth. It requires specificity. It requires the kind of observed human detail that makes a viewer feel something beyond mild recognition. The AI-generated imagery had none of this. It had the aesthetic vocabulary of emotion without any of the underlying substance. It looked like it was about something without actually being about anything.

I've shown this ad to people who work in production, in creative, in strategy. The reaction is remarkably consistent. They don't talk about the technical flaws first. They talk about how it made them feel. "Empty" comes up a lot. "Hollow." "Like a greeting card written by someone who's never received one." The technology faithfully reproduced the surface of advertising without accessing anything beneath it.

The craft argument nobody wanted to make

For the last two years, the advertising industry has been having a loud, anxious conversation about AI replacing human creative work. The arguments tend to fall into two camps. Camp one says AI will democratize creativity and make production cheaper and faster. Camp two says AI will destroy jobs and degrade quality. Both camps have been arguing in the abstract. The Toys 'R' Us ad moved the argument into the concrete, and the concrete was not kind to camp one.

Here's what the ad actually demonstrated. You can use AI to generate imagery that resembles advertising. The output will have the composition, color grading, and general visual language of professional commercial work. It will look, at a distance, like something a production company made. But the distance matters. The moment you look closely, the moment you engage with it as a viewer rather than a thumbnail, the absence of human decision-making becomes apparent. Every frame of a traditionally produced commercial reflects thousands of small choices made by people who understand what they're trying to communicate. The casting director who chose that specific child's face. The DP who placed the light at that exact angle. The editor who held that shot for an extra beat because the emotion needed room. AI doesn't make these choices. It makes approximations of these choices based on patterns in its training data.

What the brand got wrong

I have some sympathy for Toys 'R' Us here. They're a brand in a difficult position. Emerging from bankruptcy with limited budgets and a need to generate attention. Using Sora was, at minimum, a way to get people talking. And it worked. The problem is that the conversation was about the wrong thing. Nobody talked about the brand. Nobody talked about the story. Nobody talked about what Toys 'R' Us means to them or their children. Everybody talked about AI. The medium consumed the message entirely.

This is a brand strategy failure more than a production one. The decision to use AI-generated video for a brand reintroduction suggests that someone in the room believed the novelty of the technology would be a net positive for the brand. That the "made with AI" angle would generate excitement and position Toys 'R' Us as forward-thinking. Instead, it generated skepticism and positioned the brand as a cautionary tale. The ad became evidence in someone else's argument.

When you're reintroducing a beloved brand to the public, the single most important thing you can communicate is care. That someone cares enough about this brand, this story, this audience to invest real attention in getting it right. An AI-generated ad communicates the opposite. It communicates efficiency. Speed. Cost savings. Those might be real advantages in other contexts. For a brand trying to rebuild emotional connection, they're liabilities.

The argument for human craft

I don't think AI is useless in creative production. I think it will become a powerful tool for ideation, for rough concepting, for generating reference material, for accelerating parts of the production process that are genuinely mechanical. I've seen it used well in those contexts. What I don't think it can do, at least not yet and possibly not ever, is replace the human judgment that separates competent advertising from advertising that actually moves people.

The Toys 'R' Us ad is useful because it makes this argument viscerally. You don't need to be a production expert or a creative director to feel what's wrong with it. Regular people, the actual audience for the commercial, felt it immediately. They didn't need the industry discourse to tell them something was off. They knew. That instinctive audience response is the most powerful data point in the entire AI-and-creativity debate. People can tell. They may not be able to articulate what's missing, but they can feel its absence.

Where this leaves us

The technology will improve. The next generation of AI video tools will produce more consistent characters, more stable environments, more convincing motion. The technical uncanny valley will narrow. But I suspect the emotional uncanny valley will persist longer than the technologists expect. Because the problem isn't fidelity. The problem is intention. A human director points the camera at a specific moment because they understand what that moment means. An AI model generates a statistically likely frame. The output may look similar. The viewer's experience of it will not be the same.

For brands considering AI-generated creative, the lesson from Toys 'R' Us is straightforward. The audience is not impressed by your production method. The audience is moved by your craft, your specificity, and your care. If AI helps you deliver more of those things, use it. If it helps you deliver less of those things faster, you've optimized for the wrong variable.

The ad proved exactly the point it didn't intend to make. Audiences know when something was made with care. And they know when it wasn't.