This weekly update shows a complicated picture for employees working with AI agents and smart computer helpers in their daily jobs.

Major Companies Push AI Training Hard

Big companies are making employees learn AI whether they want to or not. Citi bank just started a huge training program for all 175,000 workers in 80 countries. The bank wants everyone to learn how to write better "prompts" - the instructions you give to AI tools. Workers who already know AI can finish the training in 10 minutes, but beginners need about 30 minutes.

Citi isn't alone in this push. The company Accenture is actually firing employees who can't learn to use new AI technology. Walmart's CEO recently said AI will change "virtually every job" at the huge retailer. This shows how serious companies are about making AI part of everyday work.

The Problem with "Workslop"

But employees are finding that AI-generated content often creates more problems than it solves. Researchers from Stanford University and a company called BetterUp Labs discovered something they call "workslop". This is AI content that looks professional and polished but doesn't actually help get work done.

A shocking 40% of US workers said they dealt with workslop in just the last month. This low-quality AI work forces employees to spend extra time fact-checking, rewriting, and fixing what the AI produced. The researchers calculated that this costs companies about $186 per employee every month in lost productivity.

Workers are getting frustrated because they have to clean up after AI instead of focusing on their real jobs. The AI content often fills up email inboxes and meetings with information that sounds important but doesn't help anyone make decisions.

Job Security Fears vs Reality

Many employees worry that AI agents will steal their jobs, but new research suggests this isn't happening as quickly as people feared. Scientists at Yale University studied what happened to jobs since ChatGPT launched in November 2022. They found that the job market has stayed surprisingly stable over the past 33 months.

The researchers looked at whether jobs with high "AI exposure" - meaning jobs where AI could do many of the tasks - were disappearing. They found the opposite was true. The mix of different types of jobs in the economy hasn't changed much more than during past technology changes like when computers and the internet first appeared.

This matches what happened with older technologies. Computers and the internet took decades, not just months, to really change how people work. The researchers think AI will probably follow a similar slow pattern.

Companies Try to Help Employees Feel Better About AI

At a big conference called Workday Rising 2025 in San Francisco, company leaders talked about making AI work better with humans. Ashley Goldsmith, who runs people operations at Workday, said "The human is always at the center, and AI is here to help us maximize our potential".

Workday created something called "Everyday AI" to help employees feel more comfortable using AI tools. The company wants workers to explore AI "without fear" and know they have "permission and encouragement" to try new things.

The results look promising. Workday employees who participate in AI projects are 42% more likely to get promoted within the company and 30% less likely to quit their jobs. The company also uses AI to match workers with new opportunities that fit their interests.

Trust Problems with AI Tools

Despite company efforts, many employees still don't trust AI agents to do good work. A study of British workers found that 64% think AI tools are unreliable. More than half said AI gives wrong answers but acts very confident about them. Workers also complained that AI tools ignore their feedback completely.

Some employees are so worried about AI that they hide when they use it. A survey for The Guardian newspaper found that British workers keep their AI use secret because they fear it might damage their reputation at work. This creates a strange situation where companies want employees to use AI, but workers are afraid to admit when they do.

Looking Ahead

The picture for employees and AI agents remains mixed. While companies are investing heavily in AI training and tools, workers are dealing with real problems like low-quality AI output and trust issues. The good news is that widespread job losses haven't happened yet, giving employees and companies time to figure out how to work together with AI in better ways.