Companies around the world are rushing to adopt AI agents and discovering that this technology brings both major benefits and serious risks for their workforce. This weekly update examines how businesses are navigating these changes.

A comprehensive survey by the EY organization released this week provides important insights into how companies are managing AI in the workplace. The survey included 975 top executives from 21 countries and found that organizations using more advanced AI governance are pulling ahead of competitors. Nearly 81% of companies reported improved innovation, while 79% saw better efficiency and productivity. About half of businesses experienced revenue growth (54%) and cost savings (48%).

However, the same survey revealed a darker side to AI adoption. Almost every company surveyed (99%) reported financial losses from AI-related risks. These losses averaged about $4.4 million per company, with nearly two-thirds losing more than $1 million. The most common problems included not following AI regulations (57%), hurting sustainability goals (55%), and AI systems producing biased results (53%).

One of the biggest challenges businesses face involves "citizen developers" - regular employees who create or use AI agents on their own. Two-thirds of surveyed companies allow this activity, but only 60% provide proper rules and guidelines to ensure these AI agents are used responsibly. Even more concerning, half of these companies admitted they cannot see or track how employees are using AI agents. As agentic AI becomes more common in workplaces, experts warn that risks will only grow.

Researchers from Stanford University and a company called BetterUp discovered that workers are splitting into two distinct groups when it comes to AI usage. They call these groups "pilots" and "passengers". Pilots use AI tools to enhance their creativity and make their work more precise. They use AI 75% more often at work than passengers and 95% more often outside of work. Passengers, on the other hand, rely on AI to do their work for them, often just copying and pasting AI responses without thinking.

This difference matters because passengers are creating a major problem called "workslop". This term describes content that looks professional and polished but contains little real value. About 40% of workers said they encountered workslop in the past month, and roughly 15% of workplace content now falls into this category. When people receive workslop, more than half feel annoyed, confused, or even offended. They also start viewing the person who sent it as less creative, capable, and reliable.

The financial cost of workslop is substantial. Researchers estimate it creates an "invisible tax" of $186 per employee every month. For a company with 10,000 workers, this adds up to more than $9 million annually. Over time, workslop erodes trust between coworkers, with one-third of people telling their teammates when it happens and becoming less willing to work with that person again.

The impact on employment is becoming increasingly clear and concerning. Klarna, a financial technology company in Sweden, provides a stark example. The company's CEO, Sebastian Siemiatkowski, revealed that Klarna reduced its workforce from 7,400 employees to just 3,000 after implementing AI systems. The CEO criticized other technology leaders for "sugarcoating" how badly AI will affect jobs. He specifically mentioned that thousands of translators in Brussels could be replaced by AI that already does their work.

For young people entering the workforce, the situation is particularly difficult. Many companies have stopped hiring entry-level workers because AI can now handle tasks that used to be perfect for new graduates. The Big Four accounting firms cut their graduate hiring by between 6% and 29% in a single year. This creates a serious problem: if companies do not hire and train young workers today, there will be nobody with experience to replace older workers when they retire.

In the United States, about 21% of workers now use AI in their jobs according to research from Pew Research Center. This number increased from 16% just one year ago. Workers younger than 50 and those with college degrees are most likely to use AI in their work. However, most American workers (65%) still say they do not use AI much or at all in their jobs.

Businesses face a critical choice in the coming months. They can view AI as simply a way to cut costs and reduce headcount, or they can use it as a tool to help workers become more productive and creative. The evidence from this week suggests that companies taking the second approach - with proper oversight, training, and governance - are seeing better results. However, organizations that rush into AI adoption without clear strategies risk losing money, damaging employee morale, and creating long-term workforce problems they cannot easily fix.