A recent study from EY, a large business company, surveyed over 1,100 desk workers at very large companies in the United States. These workers came from six different industries: banking, stocks and money management, stores that sell things, manufacturing, oil and gas companies, and technology companies. The study asked these workers what they think about AI agents - computer programs that work by themselves to do multiple tasks.

The study found something very interesting - workers are both excited AND scared at the same time. Eighty-four percent of workers said they were eager to use AI agents in their jobs and believe it will help them work better. These workers think AI agents will make them more productive, work more efficiently, and even make their jobs more enjoyable. But here is the problem: 56% of those same workers are very worried that AI agents will take away their jobs. Fifty-one percent even worry their entire job could become obsolete and they won't be needed anymore. This shows a big disconnect between excitement and fear.

Different workers feel different levels of worry depending on their jobs. Regular desk workers (who do not manage other people) are much more scared than managers. Sixty-five percent of workers who are not managers worry about losing their job to AI agents. But only 48% of managers have this same worry. This big difference shows that people leading companies might not understand how scared regular workers are about what AI means for their future.

Companies are not helping workers learn about AI agents the way they should. Even though eighty-six percent of workers who already use AI agents say it has helped their teams work better and faster, most workers don't understand how AI agents work. Eighty-five percent of desk workers are learning about AI agents by themselves, after work hours. They are learning at home instead of at work - companies are not training them during paid work time. Eighty-three percent of workers said that almost everything they know about AI agents is completely self-taught. This is not fair. Workers deserve to learn these important skills during their work day with help from their companies.

But here is some good news: workers who already use AI agents say the technology actually works and helps them. Eighty-six percent of workers reported that using AI agents made their teams more productive and helped get more work done faster. Ninety percent of workers who use AI agents feel very confident they can work with them. This proves something important: workers are not afraid of the technology itself - they are most afraid of losing their jobs and not having support from their companies.

The worker's worry about AI and job loss is starting to become real in the real world. In October 2025, several big technology companies announced major layoffs. Rivian, a company that builds electric cars in the United States, cut 600 workers (about 4% of its workers). This was actually the third time Rivian laid off workers this year. Meta, the company that owns Facebook and Instagram, said it will lay off about 600 workers from its AI departments. Applied Materials, a company that makes computer chip parts, will cut about 1,400 workers (which is 4% of its workforce). These real layoffs show that AI is already changing how companies hire and keep workers.

But one of the smartest AI experts in the world says workers should not panic right now. Andrej Karpathy, who helped start OpenAI (the company that made ChatGPT), said that AI agents are still far from perfect. He explained that AI agents cannot think as well as humans, cannot use computers the way we do, cannot remember things people tell them, and cannot keep learning by themselves. Karpathy said it will take about ten more years before AI agents can work without a human watching and guiding them. Right now, you should think of AI agents like hiring a new worker or an intern that you have to supervise.

The real world already shows that AI is not ready to take everyone's jobs. Ninety-five percent of companies that tried AI test projects have failed. Plus, many companies that planned to cut customer service workers by 80% using AI have changed their minds and canceled those plans. This proves the technology does not work as well as people hoped. One example: McKinsey built an AI agent that helped check incoming emails, but a human still had to check the AI agent's work before sending it out. Even with AI help, the company still needs workers.

The next ten years will be very important for workers, companies, and AI together. Companies must talk clearly with workers about how AI will change their jobs. Workers need proper training during work time, not having to teach themselves at night. Leaders need to help workers feel less scared and more confident about their future. Workers should start learning about AI now, but they also deserve real help and support from their companies. The future of work will need both humans and AI agents working together, not AI agents replacing humans.