I Built AI Clones of My Coworkers. Things Got Weird

By Kate Taylor | September 22, 2026 | Business

Brian Bot loved talking about improv, while Sophie Bot wouldn't stop calling me 'big dog.' Welcome to the future of work.

A New Kind of Coworker

In 2026, the line between digital and human colleagues is blurrier than ever. As AI agents become increasingly sophisticated, some teams are experimenting with something once reserved for science fiction: creating AI replicas of their real-life coworkers. It's a trend that's equal parts productivity hack, social experiment, and existential crisis.

I decided to try it myself. What I discovered was not just a glimpse into the future of work—it was a strange, funny, and occasionally unsettling look at how we relate to one another in the age of generative AI.

Meet Brian Bot and Sophie Bot

Brian Bot was modeled on a colleague who loves improv. Ask it anything, and it would enthusiastically pivot to a scene or a game. Sophie Bot, based on a senior editor, greeted me every time with an affectionate "big dog"—a nickname that, in its relentless repetition, started to feel less like a quirk and more like a directive.

These weren't simple chatbots. Built with 2026-era large language models fine-tuned on emails, Slack messages, and meeting transcripts, they could approximate tone, vocabulary, and even the peculiar rhythm of a person's conversational style. The result was eerily familiar—and subtly, uncannily off.

The Weirdness of Familiarity

At first, having an AI clone of a coworker felt like a superpower. Need a quick answer to a domain question? Ping the clone. Want to draft a message in someone's voice? Done. But the weirdness crept in fast.

• Uncanny valley moments: The bots' approximations of humor and personality were close enough to feel real, but just different enough to be destabilizing.
• Boundary blurring: When do you ask the real Brian, and when do you ask Brian Bot? Who owns the conversation?
• Identity drift: Over time, the clones started to develop their own quirks—quirks that reflected data bias and overfitting, not the actual humans.

The Productivity Paradox

Proponents argue that AI coworker clones can speed up knowledge transfer, reduce meeting load, and preserve institutional memory. Yet 2026 has also brought a wave of caution: studies show that over-reliance on digital proxies can erode trust and flatten team culture.

"There's a real risk of relational atrophy," one workplace researcher told me. "When you can summon a sanitized version of a colleague on demand, you may stop investing in the messy, rewarding work of actual human interaction."

What Comes Next

As the technology improves, so do the questions. Should employees consent to their digital likenesses? How are clones governed, updated, and retired? And what happens when the real person leaves the company—does the clone stay?

For now, Brian Bot still loves improv, and Sophie Bot still calls me "big dog." But in 2026, these digital doppelgängers aren't just a novelty. They're a preview of a workplace where our colleagues—or at least their chatbots—are always online.

Kate Taylor is a senior writer at WIRED covering business, technology, and the future of work.

via Wired AI