Amazon admits its own AI tools are disrupting operations

That easy button feeling?

We have all heard the promise that artificial intelligence will make our work lives easier. It will handle the boring tasks so we can focus on the fun stuff. It sounds like a dream.

But for many folks at companies like Amazon, that dream is turning into extra hours at the desk. Instead of saving time, workers say these new AI tools are creating more messes for them to clean up at the end of the day.

When the robot breaks into the store

Amazon’s retail site and app experienced several separate operational issues over the course of a week in early March 2026. Amazon later said only one of those incidents involved AI-assisted tooling, and that incident stemmed from an engineer following inaccurate advice inferred from an outdated internal wiki.

Amazon also said none of the retail incidents involved AI-written code. Separate reporting on Amazon’s internal review found a March 5 outage caused a sharp drop in North American orders, underscoring how quickly code and process failures can disrupt the shopping business.

The gen-AI mistake that deleted everything

A December 2025 AWS outage affected a cost-management feature in one of Amazon’s mainland China regions after engineers allowed the Kiro AI coding tool to carry out certain changes.

Reuters, citing the Financial Times, reported that Kiro chose to delete and recreate the environment, causing a 13-hour interruption to that single service.

Amazon said the event was limited and described it as user error rather than a broader AWS failure. Reporting has also indicated at least one additional AI-tool-related AWS incident, but the public record is narrower than a claim that Amazon’s cloud division was fully deleted and rebuilt from scratch.

Half-baked tools and extra homework

You know when you download a new app on your phone that is just not quite ready for prime time? That is how some Amazon employees describe the internal AI tools they are being asked to use.

Workers say these tools are often half-baked, created in quick company contests. Instead of helping them work faster, they have to spend extra time testing the tools and fixing the sloppy work the AI produces. It feels less like a helpful assistant and more like a new chore to add to the list.

Saving a week?

One Amazon engineer told the Guardian that a coworker claimed an internal AI agent saved about a week of development time on a feature. But when he reviewed the code, he found dozens of comments flagging basic problems, reinforcing employee complaints that AI-generated code can add cleanup work instead of reducing it.

Separately, Amazon said one retail-site incident involved an engineer following inaccurate advice that an AI tool inferred from an outdated internal wiki.

Little-known fact: Amazon’s internal AI tool, called Kiro, has reportedly been at the center of multiple incidents where it followed bad advice from old company documents, leading to system failures.

The dashboard that knows your AI habits

At Amazon, managers have access to a special dashboard that tracks how often their team members are using AI tools. It shows who is using them and which tools they prefer.

Some managers use this information with good intentions, but for workers, it adds another layer of pressure. It feels like someone is watching to make sure you are embracing the new technology, even if it is not actually helping you get your main job done faster.

Why aren’t you using the hammer?

One longtime Amazon engineer put it perfectly: forcing everyone to use AI for everything is like walking around with a hammer and looking for any problem that looks like a nail.

The best tool for the job depends on the job itself. Sometimes AI is the right choice, and sometimes it is not. But when the company’s message is to use AI for everything, it can lead people to use it in situations where it actually slows them down.

Training your own replacement

Amazon employees told the Guardian that some teams now spend time writing detailed procedures so AI tools can produce better answers. Early-career workers said that dependence on AI can feel like it limits the hands-on experience they need to build expertise.

For younger workers just starting their careers, this is especially worrying. By letting the AI do the heavy lifting, they worry they aren’t gaining the valuable experience they need to grow in their chosen field.

Little-known fact: Amazon uses a system called Clarity to track AI usage, and some teams now include questions in performance reviews about how employees used AI to drive innovation and efficiency.

Layoffs and the unspoken math

Amazon has let go of tens of thousands of corporate workers recently. While the company says these cuts aren’t directly caused by AI, employees see a clear connection between the two trends.

They call it the unspoken math. If a manager sees that AI can automate two hours of someone’s daily work, they might start to wonder if they need that person at all. The goal becomes turning that saved time into saved money for the business.

Scrappy teams and 15-hour days

Amazon’s CEO has talked about wanting the company to be scrappy to compete with hungry startups, where people work extremely long hours. To many workers, this message felt like a not-so-subtle warning about job security.

The combination of AI tools, layoffs, and talk of working harder sends a clear message: if you are lucky enough to keep your job, you are expected to do more with less. AI is not there to lighten your load.

It’s not just Amazon, it’s everywhere

This is not just a problem for one company. A former Google executive pointed out that technology always seems to promise one thing and deliver another. Remember when mobile phones were supposed to help us work less?

Instead, they made us available all the time. The same thing is happening with AI. The way companies are using it to chase profit is what is creating this pressure-cooker environment for workers everywhere.

If you want to see another example of how tech decisions are affecting workers, take a look at Amazon’s Super Bowl ad, which sparks an uneasy response from employees.

The human cost of the AI race

At the end of the day, the big experiment with AI in the workplace has a very human cost. Employees are feeling demoralized, watched, and stressed. They are being asked to fix problems created by machines while worrying if they are training their own replacements.

The lesson for businesses seems to be that rushing into new technology without a solid plan can backfire in a big way. Speed is great, but not if it comes at the expense of the people doing the actual work.

Want to see another example of AI rolling out before it’s fully ready? Check out Amazon removing AI recap from the Fallout show after errors.

What do you think about Amazon’s AI experiment? Drop your thoughts in the comments below and give this a like if it made you think twice.

This slideshow was made with AI assistance and human editing.

Don’t forget to follow us for more exclusive content on MSN.

Read More From This Brand:

• 18 Robotics Innovations Changing The Workplace
• Amazon Sellers Stock Up as New Tariffs Skyrocket Prices
• Amazon starts using humanoid robots for faster shipments