AI may be the future, but for many workers it’s currently creating more headaches than help. And there’s a new term for it — “workslop”. Inspired by research from Stanford and BetterUp Labs featured in Harvard Business Review, workslop refers to AI-generated output that looks like productive work but is actually useless. It’s jargon-filled, lacking context, and ultimately unhelpful.

From slick PowerPoints to official-looking reports or realistic code, AI content can seem legit at first glance. But once it lands in the hands of professionals, it often falls apart. Instead of saving time, it creates more work — checking, correcting, and redoing. One retail director said they had to research the facts, hold meetings, and redo the project themselves after receiving AI-generated material.

The numbers paint a clear picture. Forty percent of surveyed workers said they had received workslop recently. Each instance costs nearly two hours of someone’s time. That adds up to about $186 per employee per month, or over $9 million annually for a 10,000-person company, all spent cleaning up bad AI output.

Despite the hype, AI tools are not delivering meaningful productivity gains. MIT research shows they aren’t increasing revenue either. Meanwhile, workers are told to use AI, but also to double-check and fix everything it produces. It’s a lose-lose situation, and the costs keep rising.

AI was supposed to transform the workplace. So far, it’s mostly producing polished nonsense. Or as we now call it — workslop.