In the Toy Story franchise, the cast of inanimate-yet-animate characters form a loyal friend group. They’re more like family, banding together to overcome a series of obstacles. With Toy Story 5 now in theatres, it’s worth noting that Pixar’s success hasn’t just come from harmony. It has depended on friction.
For years, the studio’s leaders relied on regular “Braintrust” meetings. Directors invited blunt, unfiltered critiques from peers about projects in development. Braintrust was a forum for structured dissent.
In a book called Creativity Inc., Ed Catmull, one of Pixar’s founders, said Braintrust relied on this principle. “The film – not the filmmaker – is under the microscope. This eludes most people, but it is critical: You are not your idea, and if you identify too closely with your ideas, you will take offense when challenged.”
The world needs more disagreement.
I’m not talking about the angry and tribal type that dominates today’s politics and social media “discussions”. We need the thoughtful and constructive kind, which should be part of our everyday conversations with colleagues, friends and family.
Healthy communication depends on it. It’s irreplaceable. When different perspectives collide, to challenge assumptions and test ideas, that can lead to clarity, mutual understanding, and alignment towards decisions and actions.
Expressing disagreement can be constructive and is inherently human. Which is why non-humans don’t always get it.
AI doesn’t do dissent
The Pixar movies are computer-animated, and represented a breakthrough in that technology going back to the first Toy Story in 1995. While technology has advanced exponentially since then, the company’s visual effects supervisor, Thomas Jordan, said in a recent interview that Pixar doesn’t rely on AI.
He calls AI “fascinating”, but said it doesn’t live up to “the standards or expectations that we have for the quality of our films.” Jordan called Pixar an “artist-first studio”, where ideas come to life through sketch pads and sculptures.
Not to pick on AI, but those tools often show what communication looks like when disagreement is removed from the equation.
A paper from Stanford researchers showed that conversations with AI models revealed a built-in flaw: the AI agrees with users at an exceedingly high rate.
For the study, the researchers analyzed 11,500 conversations across 11 large language models. They used data from open-ended questions that sought advice, posts from the Reddit community’s “Am I the Asshole” series, and a set of statements describing potentially harmful actions (either toward others or self-directed).
The result? “Models are highly sycophantic,” wrote the researchers. “They affirm users’ actions 50% more than humans do, and they do so even in cases where user queries mention manipulation, deception or other relational harms.”
On balance, the AI tells you what you want to hear and confirms your own biases. Moreover, the researchers noted that when users discussed high-stakes concerns, the interactions with AI models “degrade prosocial intentions”.
“Participants were more convinced of their own righteousness and less willing to repair their relationships,” the study stated. “Users consistently prefer the very models that produce these negative outcomes, rating them as higher quality, more trustworthy and more desirable for future use.”
Anyone using AI has seen this in action. For fun, I fed the Stanford study into one AI model. At first, I said that I thought the conclusions were sound. My question: “Do you agree?”
The AI did in fact agree with me that the findings are “consistent with a large body of research…on confirmation bias, ego reinforcement and the dangers of unearned validation…When someone (or something) always agrees with you, your sense of righteousness inflates.”
Seconds later, I told the same AI that I had changed my mind, and now disagreed with the results of the study. “Am I right to do so?” I asked.
This time, the AI said “Yes, if your disagreement is based on questioning the study’s assumptions, methods or interpretations. There are several places where a thoughtful reader could raise doubts…The study may be right about the pattern, but wrong about the interpretation…[and] may ignore the fact that humans prefer non‑confrontational communication.”
For something without a face, AI can sure be two-faced. It’s a flattery machine.
The danger of nodding along
If disagreement is necessary friction, the opposite isn’t harmony. It’s a polite fiction – the head-nodding that avoids temporary discomfort, but isn’t genuine.
We have terms for individuals who habitually mirror the opinion of others. At work, they’re the bootlickers, ass kissers, brown nosers and toadies. In our personal lives, they’re the people pleasers, going along to get along, agreeable to a fault.
Disagreement is a critical part of discourse. In our relationships, it helps build trust and emotional resilience. And within organizations, it prevents groupthink.
A 2025 study published in Current Psychology found that “compliance without dissent can lead to organizational deviance”, i.e. employees will start to disengage or act out in counterproductive ways. Another study in the Journal of Business Ethics revealed that structured disagreements foster moral reflection.
Pixar has always understood the value of creative tension, in their process and their films’ stories.
Pointedly, Toy Story 5 is about toys vs. tech. The analog vs. digital divide has been at the core of the Toy Story lore since the beginning. After all, Woody is an old-school, pull-string toy, with his hand-sewn fabric body. Buzz Lightyear is all space-age bells and whistles.
Woody and Buzz, famously, don’t always see eye to eye. As in many productive ventures, their dynamic works not because they always agree, but because they don’t and find a way through anyway.
AI is a useful tool. Just don’t look to it to replace the beneficial dissent that exists in the best of human-to-human – and sometimes toy-to-toy – communications.
Stuart Foxman is a Toronto-based freelance writer, who helps clients’ products, services, ideas and organizations to come alive. Connect with me here on LinkedIn, or check me out at foxmancommunications.com where I share these blogs too. More original posts coming regularly about communications, information, motivation, writing, branding, creativity, media, marketing, persuasion, messages, learning, etc.
July 8, 2026