TL;DR
A developer on GitHub has released a PowerShell project called Remove Windows AI that aims to strip AI components from Windows 11. The repo and observers warn the script can trigger antivirus alerts and recommend testing in a VM; prominent critics of desktop AI have publicly supported the effort.
What happened
A GitHub user who goes by "zoicware" published a PowerShell project named Remove Windows AI that its authors say is designed to eliminate AI-related components from Windows 11 builds, including 25H2 and future releases. The repo, worked on since 2024 with contributions from developers identified as "adeel26in" and "csmit195," positions the script as a way to improve user experience, privacy and security by removing what it describes as increasingly pervasive AI features. The project’s maintainers caution that third‑party antivirus products may flag the code as malicious and advise users who are uncertain to run it inside a virtual machine. The Register recommended only people capable of reviewing the scripts’ code should consider running it. The effort has drawn attention on social media, including praise from Meredith Whittaker, president of Signal, who framed it as community harm‑reduction against entrenched AI integration in Windows.
Why it matters
- Raises questions about user control over AI features bundled into mainstream desktop operating systems.
- Highlights privacy and security concerns tied to new OS-integrated AI components that some critics say create new data exposures.
- Demonstrates community pushback and the emergence of user-created tools to modify or remove vendor-supplied functionality.
- Signals potential risks for nontechnical users who may try such scripts despite warnings about antivirus flags and the need to review code.
Key facts
- Project name: Remove Windows AI, hosted on GitHub by user "zoicware".
- Repo contributors include developers identified as "adeel26in" and "csmit195" and work began in 2024.
- The script targets AI features present in Windows 11 25H2 and future builds, according to the repo description.
- Maintainers warn third‑party antivirus packages may falsely detect the script as malicious and recommend VM testing for uncertain users.
- The Register advised only people able to inspect the code should run such scripts.
- Meredith Whittaker, president of secure messaging service Signal, promoted the project on BlueSky, calling it community harm reduction.
- Whittaker has previously criticized Microsoft Recall and warned about insecure AI-enabled features at the 39th Chaos Communication Congress.
- The source cites broader objections to AI including privacy, training-data legality, environmental cost, bias, misinformation and a July 2025 review finding no robust link between AI adoption and aggregate productivity gains.
What to watch next
- Whether Microsoft responds formally to community tools that remove or alter bundled AI features — not confirmed in the source.
- If similar projects proliferate to give users more control over AI components in other desktop environments — not confirmed in the source.
- Whether evidence emerges validating the specific security and privacy claims made about Windows AI features and the script's effectiveness — not confirmed in the source.
Quick glossary
- PowerShell script: A sequence of commands and instructions written for Microsoft's PowerShell shell and scripting language, used to automate administrative tasks on Windows.
- GitHub: A web-based platform for hosting and collaborating on code repositories, often used to publish open-source projects.
- AI agent (agentic AI): A form of AI designed to perform tasks autonomously or with some degree of decision-making, as opposed to providing only passive outputs.
- Virtual machine (VM): A software-based emulation of a computer system used to run separate operating systems or test code in an isolated environment.
Reader FAQ
Does the script remove all AI features from Windows?
The project says it aims to remove all AI features from Windows 11 builds including 25H2, per the repo description.
Is the script safe to run on a personal machine?
The repo warns that third‑party antivirus may flag it and recommends testing in a virtual machine; the Register suggests only people who can review the code should run it.
Has Microsoft endorsed or commented on the script?
not confirmed in the source
Why are some people opposed to AI features in Windows?
Criticisms cited include privacy and security risks, the use of scraped training data, environmental impact of data centers, bias and errors in outputs, and effects on public discourse and skills development.
OSES Developer writes script to throw AI out of Windows Satya Nadella's call to accept and embrace desktop brainboxes faces skepticism Thomas Claburn Tue 13 Jan 2026 // 07:23 UTC Software developers have created…
Sources
Related posts
- A deep dive on agent sandboxes — Codex, macOS Seatbelt, Linux seccomp and policy
- Meta-backed Hupo pivots from mental wellness to AI sales coaching, raises $10M
- Google pulls some AI health summaries after probe finds dangerous errors