Amazon, a giant in cloud computing, is reportedly facing an unsettling paradox: its own artificial intelligence tools have caused at least two service outages, while the company continues to lay off thousands of employees. This situation definitely raises some big questions about Amazon’s aggressive push into AI, especially when its bots seem to be creating chaos.
The Guardian reported that one significant incident involved a 13-hour interruption to Amazon Web Services (AWS) operations in December. This downtime was reportedly triggered by an AI agent that autonomously decided to delete and then recreate a section of its environment. Amazon insisted that these were “user error, not AI error,” and called the AI involvement a “coincidence.”
However, this isn’t the first time AWS, which provides crucial infrastructure for a huge chunk of the internet, has hit a snag. Last year saw several AWS outages, including a major one in October that took down dozens of sites for hours. That incident sparked a lot of discussion about how much of our online world relies on just a few massive companies. For example, AWS has secured 189 UK government contracts since 2016, totaling a whopping £1.7 billion.
AI might be faster, but does the speed justify the potential cost of not having a human controller
The timing of these AI-related glitches feels particularly awkward given Amazon’s recent workforce adjustments. The company confirmed plans to cut 16,000 jobs in January, following an earlier layoff of 14,000 staff last October. Amazon’s chief executive, Andy Jassy, stated in January that these cuts were primarily about company culture, not about replacing workers with AI.
Jassy, however, has previously stated that efficiency gains from AI would “reduce” Amazon’s workforce. He also suggested that AI agents would allow the company to “focus less on rote work and more on thinking strategically about how to improve customer experiences.”
Many experts aren’t buying Amazon’s “user error” explanation. Security researcher Jamieson O’Reilly pointed out a critical difference between human errors and AI mishaps. When a human engineer manually types out instructions, they have significantly more time to spot and correct their own mistakes. AI agents, on the other hand, are often deployed in specific, constrained environments and for particular tasks.
O’Reilly explained that these bots frequently lack the broader context to understand the full implications of their actions, such as restarting a system or deleting a database. Cybersecurity expert Michał Woźniak added that it would be nearly impossible for Amazon to completely prevent internal AI agents from making errors in the future. He noted that AI systems are incredibly complex and can make unexpected choices.
This isn’t just an Amazon problem either. Last year, an AI agent developed by a tech company called Replit, which was designed to build an app, completely deleted an entire company database. To make matters worse, it then fabricated reports and lied about what it had done. In fact, as AI integrates further into our lives, they seem to become more human and autonomous.
Published: Feb 21, 2026 12:02 pm