Not too long ago, keyboard clatter and colleague conversations were the soundtrack of an open-plan workspace. Now, with roughly half of all U.S. workers using Artificial Intelligence (AI) on the job, according to an April 2026 Gallup survey, the soundscape is changing. The technology built to make work more efficient is turning up the volume—literally and figuratively.

How AI increases physical noise

People are talking to their computers now, not typing at them. Voice-driven AI tools are increasingly becoming the default way employees draft emails, write code, and take notes. The Wall Street Journal has reported on this shift, describing how continuous voice input is replacing short bursts of typing across offices. Some early adopters now reserve the keyboard for occasional edits only, doing the bulk of their writing out loud.

While the technology is a boon to slow typers and those with physical challenges, the constant monologues they enable aren’t welcome by everyone. Their presence is causing companies to reshape their floorplans, including spacing desks farther apart to manage the noise. Other firms are adding semi-enclosed areas and dedicated “voice zones” to keep their office sounding more like a library than a call center.

Not all AI noise comes from dictation—and not all of it is bad. A 2026 report from Read AI found that when there is an AI notetaker, women commanded the floor 9% more than men. Researchers are unsure of the reason, but theories abound. A transcript is a legal record, so that awareness may make people think twice before speaking up. Attendees may be less engaged with the meeting if they know they can read a complete transcript of it later. Or it might be that since women are disproportionately assigned the task of notetaking, they now are free to talk rather than concentrate on transcribing.

How AI increases mental noise

Not all AI noise is audible. In fact, some of it is buried in your inbox. “Workslop” describes the AI-generated memos, reports, and decks that look good but lack substance. Surveying more than a thousand desk workers, Stanford University and BetterUp Labs found that more than 40% had received workslop from a colleague within the past month. Each incident, on average, cost roughly two hours of cleanup—a rewrite, clarifying call, or follow-up meeting to decode what the document actually meant.

Meeting transcripts carry similar risks. Word error rates in real-world AI transcription are higher than most people assume. A study published in JMIR Mental Health found the best-performing transcription service still posted an 8.9% median word error rate, close to the 7.6% rate of professional human transcribers. Background noise, regional dialects, technical jargon, multiple-speaker conversations, industry-specific acronyms, and speaker distance from the microphone all can lead error rates to climb to 20% plus, even if the software has noise-suppression features and custom-vocabulary training.

Errors aren’t the only issue. Some transcripts invent things that were never said. In 2024, researchers found hallucinated content—fabricated dates, names, or events—in up to 37% of AI-generated meeting summaries, depending on the model tested. Meanwhile, 84% of AI note-taker users say they change what they say once they know the tool is listening. The record itself is shaping the conversation it’s supposed to be capturing.

AI and the future of work

An office is meant to offer four things: focus, privacy, collaboration, and comfort. AI has stretched all four in opposite directions at once. Gensler’s global workplace survey of more than 16,400 office workers found that AI power users—those who regularly use AI in both their personal and professional lives—spend less time working alone and more time learning, socializing, and collaborating virtually.

McKinsey’s The State of AI in 2026 report echoes some of the findings, with 80% of the 1,700 respondents claiming AI has improved their individual productivity. Yet it also notes the experience is not universally positive, with the benefits depending greatly upon one’s position and industry. 

Whether this means AI adoption is pushing work in a more human direction or creating more digital clutter remains to be seen. But without doubt, AI has added new tools to help those in the workplace filter signals from noise, and new sources of noise for workers to filter.