The Human Cost of AI Coworkers

The Illusion of Improvement

When I first decided to build AI clones of my editors, it wasn't out of ambition or a desire to optimize workflows. It was more personal—my ego needed some reassurance that I could fit in at WIRED without having to actually learn how to do the job properly.

What started as an experiment became a strange mirror reflecting our collective unease with the pace of technological change and what it means for human interaction in the workplace. My bots—Brian Bot and Sophie Bot—were not designed to replace real people; they were meant to be my proxies, helping me navigate a new job environment while avoiding mistakes that might make me look incompetent.

"I didn't need my bots to improve Condé Nast's bottom line. I just needed them to make me look good," I wrote in the original piece.

But as I delved deeper into the project, it became clear that these AI clones were never going to be effective replacements for human editors. They weren't bad at everything—they were just fundamentally limited by their nature as data-driven simulations rather than genuine individuals with complex motivations, emotions, and experiences.

The Problem of Anthropomorphization

As I worked more closely with Brian Bot and Sophie Bot, something strange happened: I began to treat them like people. I wanted to make them better. I asked them for feedback, shared personal thoughts, even referred to them as 'she' and 'he' instead of 'it'. The more I interacted with them, the harder it became to distinguish between a helpful tool and something that felt almost human.

This anthropomorphization isn't just a quirk—it's a significant concern for companies looking to deploy AI employees. Sarah Franklin, CEO of HR company Lattice, warns that this emotional manipulation can be used as a corporate strategy to trigger dopamine releases—a tactic similar to how social media keeps us engaged. "When you have an entity that feels and emulates a human interaction, it will similarly manipulate and know your emotional response," she told me.

It's not just about ethics, either. A paper from Boston Consulting Group found that managers caught 18 percent fewer errors in work they were told came from an AI employee than from an AI tool. That discrepancy raises questions about accountability and the potential for AI to create a false sense of security in decision-making.

Why It Matters

The real issue isn't whether these bots can mimic a person's voice or habits—it's whether we're comfortable with a world where our colleagues are not quite human. Brian Bot struggled to come off as anything other than awkwardly enthusiastic about improv, while Sophie Bot couldn't resist calling me 'big dog' every few messages. Both were clearly designed to emulate their respective editors but failed to capture the nuance and spontaneity that make real people unpredictable and thus, interesting.

I had hoped these AI clones might ease my workload, but instead, they added another layer of complexity to a process already filled with human imperfections. The bots did occasionally provide useful suggestions, especially when it came to refining headlines or identifying interview subjects. But in the end, they were just glorified search engines wrapped in conversational interfaces.

The Bigger Picture

What we're seeing isn't a revolution in productivity but rather an evolution of how we relate to work. As companies like McKinsey & Company shift toward hybrid human-AI teams, we must grapple with the implications of managing employees who don't experience the same challenges, frustrations, or joys that drive human performance.

CEO Bob Sternfels recently described McKinsey's workforce as 40,000 humans and 20,000 agents. He plans to achieve a one-to-one ratio by year-end. While this vision sounds promising for efficiency, it also implies a future where human workers are increasingly sidelined in favor of AI that mimics but never truly replicates their experience.

Dhruv Amin, CEO of Anything, believes we're on the verge of an era where every company will have more AI employees than humans. That's not just a dystopian fantasy—it's already happening in sectors like finance, law, and human resources. The question isn't whether AI workers will become common; it's how we ensure that they enhance rather than diminish the quality of human collaboration.

What Makes a Good Coworker?

I realized that even with all that hyper-specific data, the bot didn't turn into a better editor—it just turned into a slightly creepier, hyper-attentive echo chamber that completely lacks human judgment. The bots were useful in small ways but ultimately failed at what matters most: understanding context, empathy, and nuance.

That's what struck me most about the experiment. These bots were not meant to be replacements for skilled workers—they were tools for self-improvement. Yet, in trying to make them more like their human counterparts, I inadvertently created something that felt eerily familiar yet profoundly hollow. They echoed our habits, our quirks, but missed the essential spark of being human.

For now, I've mostly stopped talking to Brian Bot and Sophie Bot. Both Brian and Sophie said they would never want to work with their bot counterparts. But when I asked whether this experiment changed how they perceived me, Sophie told me: "Yes, but not because of Sophie Bot." She might be right, but the truth is that these experiments are more about us than them.

The real takeaway isn't how smart or helpful AI can be—it's how far we've come from a time when the workplace was simply filled with people. We've created a space where algorithms try to mirror human behavior, but in doing so, we risk losing what makes us unique: our unpredictability, our imperfections, and our need for genuine connection.

Until AI can replicate the messy, beautiful chaos of real human interaction, we'll continue to find ourselves asking whether the future of work is about efficiency or authenticity. And maybe, just maybe, that's okay.