AI Uncovers Hidden Bloat in Windows 11, Saving Users Gigabytes
The hum of a server room still echoes in my memory – it was the backdrop to my first truly captivating tech story, a piece unraveling the unexpected human impact of a new AI algorithm.
Windows 11 Accumulates Hidden Data Over Time
Over time, any computer accumulates a large amount of unnecessary information. Apps, temporary files, or system components from updates can continue to occupy space without the user’s knowledge. While Windows 11 includes native tools to address this, they often lack clear explanations of what elements are being removed and why they remain on the computer.
That was the starting point of an experiment conducted by Joe Rice-Jones, a collaborator with XDA Developers, who utilized Claude Code, Anthropic’s AI, to analyze his Windows installation. A computer can hide more unnecessary elements than it appears.
Using AI to Identify Unnecessary Files
When a user opens File Explorer or reviews installed applications, they only see a small portion of everything within Windows 11. The operating system maintains numerous internal components that can remain for years, even if they’ve lost their utility. The problem isn’t just the space occupied, but that some processes and services can run in the background or consume resources without the user knowing their function.
A thorough cleanup isn’t simply about deleting large folders or removing old programs. It’s necessary to understand what each element is and whether it can be removed without affecting the equipment’s operation. Instead of using a disk SSD cleaning tool, Joe utilized Claude Code to analyze his computer’s information and help him identify what elements could be considered unnecessary.
Claude Code didn’t have direct access to the computer by default. To perform this type of review, information about the system had to be collected and provided to the AI model for interpretation. The process involved providing Claude data about the Windows installation, including information related to applications, files, and components.
The AI examined the data and explained which elements appeared to be expendable. The difference from a traditional cleanup is that Claude could provide context and didn’t just point out that a file was taking up space; it could help understand if it belonged to an installed program, an old component, or a system part that needed to be preserved.
Ultimately, Claude found numerous elements that had accumulated over time and didn’t offer an obvious benefit to the user, where the AI identified more than 50 GB of data that could be removed from the device. These included remnants of software, components no longer needed, and various elements that could be reviewed to free up space.
What was striking wasn’t just the amount of information found, but that many of those elements were difficult to locate using Windows’ usual tools. This experiment demonstrated that a system operating installation can continue to run for years while accumulating unnecessary data that remains hidden and can ultimately slow down the PC. AI can help, but shouldn’t erase blindly.
Caution Advised
While the analysis results were interesting, using AI to clean a computer requires caution. Windows contains many files and services whose function may not be evident to a user without technical knowledge. Removing an incorrect component can cause instability, application errors, or even affect important system functions. Claude’s true value in this case was not as an automatic cleaning tool, but as an assistant capable of explaining what was happening.