GNOME bans AI-generated extensions after flood of poorly-written code overwhelms reviewers

GNOME bans AI-generated extensions

Earlier this month, the GNOME Shell Extensions store updated its review guidelines to include a new section specifically stating that "extensions must not be AI-generated," as reported by It's FOSS and Phoronix. Developers producing add-ons for the Linux desktop environment can still use AI as a tool, but if their extension's code contains signs of being mostly written by AI, the updated guidelines say it will be rejected: One of the developers who reviews extensions for Gnome, Javad Rahmatzadeh, explained in a blog post that the rule change was necessary to address an influx of extension submissions with poorly-written, AI-generated code. According to Rahmatzadeh, AI "has led to receiving packages with many unnecessary lines and bad practices," as he explains in the post: The number of submitted packages to EGO is growing every month and we see more and more people joining the extensions community to create their own extensions. Some days, I spend more than 6 hours a day reviewing over 15,000 lines of extension code and answering the community. In the past two months, we have received many new extensions on EGO. This is a good thing since it can make the extensions community grow even more, but there is one issue with some packages. Some devs are using AI without understanding the code. This has led to receiving packages with many unnecessary lines and bad practices. And once a bad practice is introduced in one package, it can create a domino effect, appearing on other extensions. That alone has increased the waiting time for all packages to be reviewed. Extensions for GNOME can serve all kinds of functions, including some that are vital to how the desktop environment runs. For instance, one of the most popular is Dash to Dock, which gives users more control and customization options for the app dock at the bottom of the screen. The new guidelines clarify that developers are still allowed to use AI as a learning aid and development tool when making their extensions. They just shouldn't be using AI to generate all of their code without understanding what it does. The Fedora Council introduced similar guidelines around vibe-coding earlier this year, although their policy wasn't a blanket ban.

New Rule Forbids GNOME Shell Extensions Made Using AI Generated Code

The GNOME.org Extensions hosting for GNOME Shell extensions will no longer accept new contributions with AI-generated code. A new rule has been added to their review guidelines to forbid AI-generated code. Due to the growing number of SNOME Shell extensions looking to appear on extensions.gnome.org that were generated using AI, it's now prohibited. The new rule in their guidelines note that AI-generated code will be explicitly rejected: "Extensions must not be AI-generated While it is not prohibited to use AI as a learning aid or a development tool (i.e. code completions), extension developers should be able to justify and explain the code they submit, within reason. Submissions with large amounts of unnecessary code, inconsistent code style, imaginary API usage, comments serving as LLM prompts, or other indications of AI-generated output will be rejected." This blog post by GNOME developer Javad Rahmatzadeh outlines the growing number of AI-generated extensions they had been seeing for inclusion on extensions.gnome.org (EGO) and thus set the new rule to forbid such AI-generated code moving forward. Users can use AI to generate their own local GNOME Shell extension but they will not be accepted to extensions.gnome.org for easy consumption by other GNOME users. This Week in GNOME outlined some new (non-AI) GNOME Shell extensions activity such as the Adaptive Brightness extension now supporting GNOME 49's new brightness slider behavior and a new All-In-One Clipboard release.

GNOME is cracking down on AI-generated code after updating its extensions guidelines

* GNOME now rejects extensions containing large amounts of AI-generated code; small AI assistance is allowed. * Unnecessary or inconsistent AI code, fake API use, or comments that are LLM prompts will be rejected. * Developers must explain their code to keep extension quality high; AI-written answers don't satisfy reviewers. Now that we've had a lot of time to play around with AI-generated code, people are asking themselves: does letting an LLM program for you improve productivity? Right now, it seems like a strange mix of "Yes" and "No." While it does allow people with zero programming skills to create an app with ease, the code itself isn't always the best it can be. While vibe coding is perfect to see how a program could behave, it's not ideal for actually shipping the product. Well, it seems that people have been taking the latter path a little too aggressively over at the GNOME extensions hub. The GNOME Project has updated its extensions guidelines to forbid huge chunks of AI-generated code, as the reviewers are unhappy with its quality. 5 GNOME tweaks that I can't live without Here are some tweaks I always make to GNOME after installing a Linux distro. Posts By Rich Edmonds Oct 10, 2025 Extensions must not be AI-generated, says GNOME A little AI is okay, though As spotted by Linuxiac, an update has arrived on the GNOME Shell Extensions Review Guidelines. Under the new section titled "Extensions must not be AI-generated," The GNOME Project explains that it's no longer interested in receiving submissions that use a large amount of AI code: While it is not prohibited to use AI as a learning aid or a development tool (i.e. code completions), extension developers should be able to justify and explain the code they submit, within reason. Submissions with large amounts of unnecessary code, inconsistent code style, imaginary API usage, comments serving as LLM prompts, or other indications of AI-generated output will be rejected. The reasoning seems very valid: the reviewers at The GNOME Project wish to enquire with extension developers about their code and the decisions they made while creating it. If the person shrugs and says, "I dunno, that's just what the AI made," then it makes the reviewer's job a lot harder. In fact, as Linuxiac further reports, people have already tried replying to queries about their AI-generated code with more AI. When the GNOME Project asks someone who submitted AI code why they wrote a specific chunk of code the way they did, they get a reply written by the same AI discussing the process. As such, to ensure developers actually know what they're submitting, the GNOME Project is rejecting code it believes is mostly written with AI. The GNOME reviewers have already noticed that any bad habits that AI introduces within the ecosystem are propagated within other projects, so demanding a high quality of work from the developers helps keep the codebase clean and efficient.