The intention of this list is to raise awareness of AI/LLM usage in popular open-source software. Provided below is an informed set of AI-free alternatives for users and developers to consider should their ethical boundaries be crossed or tolerance for risk be exceeded. This list is not a resource to be used for the harassment of other open-source developers. If you wish to advocate for the cessation of use and/or removal of AI-generated code from another project, we ask that it be done respectfully and constructively.
> This is a fork of a repo by the same name maintained by the @gen-ai-transparency org on Codeberg. If they make their repo available again, we will contribute back.
A policy that permits the use of AI/LLMs in any capacity or is declared to be [vibecoded](#vibecode). Both vibecoding and opening the door for people to vibecode count as a permissive AI policy.
Evidence can be:
- an explicit AI policy (e.g. `AI_POLICY.md`, `CONTRIBUTING.md`, developer docs) in the repo or something in the project's contributing guidelines or in their website's documentation that says that AI/LLMs are allowed
- an LLM friendly `AGENTS.md`, `CLAUDE.md`, or other such LLM instruction files or folders.
- core maintainers' blog or social media post about vibecoding
- link to readme, website, or documentation stating the project is vibecoded
> [!Important]
> If a core maintainer has noted that the code is entirely vibecoded, please put that as the *first* evidence link.
Asking an AI to write software for you (often without human review). Per [Wikipedia](https://en.wikipedia.org/wiki/Vibe_coding):
> In computer programming, vibe coding is a software development practice assisted by artificial intelligence (AI) such as by chatbots (programs that simulate conversation) or AI agents such as Codex or Claude Code. The software developer describes a project or task in a prompt to a large language model (LLM), which generates source code automatically. Vibe coding may involve accepting AI-generated code without reviewing the output thoroughly, instead relying on results and follow-up prompts to guide changes.
The term was coined by [Andrej Karpathy](https://en.wikipedia.org/wiki/Andrej_Karpathy) (co-founder of [OpenAI](https://openai.com/)) in February 2025.
A **Vibecoder** is a person who uses AI prompts to generate anything, typically code or images. It would therefore be appropriate to call Andrej Karpathy a vibecoder.
This would include the ability to enable or disable an AI feature, such as an AI assistant or AI summary feature. This would also cover the case when a project depends on an AI program or library, such as an LLM SDK.
Evidence can be:
- a link to the docs explaining the AI Functionality
- a blog post advertising the AI Functionality
- a link to the default branch in a repo that shows where the AI Functionality is implemented
An AI agent/bot is used to review pull requests, which makes slop acceptance more probable and maintainability more difficult, especially when human reviews are scarce.
Evidence can be:
- a link to an AI code review requested by a maintainer.
- an AI code review CI workflow.
- a PR where an AI code review seems to be automatically created by some sort of third party app
AI "Art" is being used in a project. This is typically for banners, avatars, promotional material, and/or blog posts. This means they're using tools that steal art from visual artists.
This means someone is asking for help with research. They would like it if you found the last known good version of a project or suggestions on alternatives for people to use. You can submit a pull request to add this info.
This is a section for repos that are similar to this one either because they are also forks of the original upstream, or because they align with our goal of identifying and avoiding AI/LLM usage OR identifying/elevating projects that do NOT use AI/LLMs. These repos are not all maintained by the same people, but we may share some contributors and are generally kind FOSS neighbors. ✨
- [llm-afflicted-software](https://codeberg.org/ai-alternatives/llm-afflicted-software) offers a similar list to ours here, however in YMAL file format for each category, making it more suitable to consumption via code (for instance writing a CLI, GUI, or TUI tool to consume it). They've also kindly reached out in the past and submitted Issues and PRs (e.g. [#12](https://codeberg.org/small-hack/open-slopware/issues/12), [#22](https://codeberg.org/small-hack/open-slopware/pulls/22)) to help us here.
- [forge.starlightnet.work/Team/No-AI/](https://forge.starlightnet.work/Team/No-AI/) which is the repo for this [no-AI list](https://noai.starlightnet.work/list.html) that lists projects that have explicitly pledged not to use AI or have an explicit policy to not use AI.
- [slop-free-index](https://codeberg.org/brib/slopfree-software-index) is a list of software that has taken steps to reject AI in its development processes.
> We've recently added a "Last Untainted Version or Commit ID" section to our below tables. This is meant for tech savvy individuals to be able to fork the project and continue maintaining it without the use of AI. Choosing to install an older version of software could expose you to to security risks over time. Only do this if you are able to accept such risks.
> * [Epiphany] (by GNOME, WebKit based) does not currently have a stance, but GNOME developers as a whole [show aware anti-"AI" sentiments](https://discourse.gnome.org/t/loupe-no-longer-allows-generative-ai-contributions/27327) and roll out [wide reaching rules](https://gjs.guide/extensions/review-guidelines/review-guidelines.html#extensions-must-not-be-ai-generated) against it.
> * [Servo] (New engine, formerly owned by Mozilla, now by a co-op) is not yet daily driver ready, but has [strong protections](https://book.servo.org/contributing/getting-started.html#ai-contributions) for when it is some time in the future.
| Searxng | [LibreY](https://github.com/Ahwxorg/librey/) | [Copilot used](https://github.com/searxng/searxng/issues?q=copilot) in reviews | Note: maintainers experimenting with and open to adding AI results, see [1](https://github.com/searxng/searxng/issues/2273), [2](https://github.com/searxng/searxng/issues/2008), [3](https://github.com/searxng/searxng/issues/2163)
| Name | Last Untainted Version or Commit ID | Tags and Evidence |
|---|:---:|---|
| [CoMaps](https://www.comaps.app/) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://codeberg.org/comaps/Governance/src/branch/main/AI_USAGE.md)) |
| [Organic Maps](https://organicmaps.app/) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/organicmaps/organicmaps/blob/master/.github/copilot-instructions.md)) |
| [OsmAnd](https://osmand.net/) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/osmandapp/OsmAnd/blob/master/AGENTS.md)) |
> [!Note]
> #### Mapping Alternatives
> * [Navit](https://github.com/navit-gps/navit) for car navigation
> * [Open Street Map](https://www.openstreetmap.org) provides lists of clients on its [wiki](https://wiki.openstreetmap.org/wiki/Software). For example, other mobile clients can be found on the [Android](https://wiki.openstreetmap.org/wiki/Android) page.
| Name | Last Untainted Version or Commit ID | Tags and Evidence | Alternative(s) |
|---|:---:|---|---|
| [Atuin](https://github.com/atuinsh/atuin) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/atuinsh/atuin/blob/main/AGENTS.md), [2](https://github.com/atuinsh/atuin/pull/2777#issuecomment-2944105696))<br />[](#ai-functionality) ([1](https://github.com/atuinsh/atuin/pull/3199)) | Built-in shell history |
| [tmux](https://github.com/tmux/tmux) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/tmux/tmux/wiki/Contributing#use-of-ai)) | [dtvm](https://www.brain-dump.org/projects/dvtm/), [mtm](https://github.com/deadpixi/mtm) |
| [Nextest](https://github.com/nextest-rs/nextest) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/nextest-rs/nextest/commit/3853279b41cc4a81f82e26fd88fac3bf85054af5)) | Just use `cargo test` |
| [Golly](https://golly.sourceforge.io/) | [](#request-for-help) | [](https://codeberg.org/small-hack/open-slopware#gen-ai-art) ([1](https://golly.sourceforge.io/Help/changes.html)) | [](#request-for-help) |
| [Just](https://just.systems/) | [](#request-for-help) | [](#ai-functionality) ([1](https://just.systems/man/en/model-context-protocol.html)) | A Make implementation |
| [Mold](https://github.com/rui314/mold) linker | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/rui314/mold/commit/dac20fa24373f35b5dec44e4740db85c3eb7b3dd)) | GNU ld |
| [Taskfile](https://taskfile.dev/) | [](#request-for-help) | [](#ai-code-reviews) ([1](https://github.com/go-task/task/pull/2592#pullrequestreview-3596720069)) | [](#request-for-help) |
| [Wild](https://github.com/wild-linker/wild) linker | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/wild-linker/wild/blob/9027437a8776afea1a1f0840fcda3fd6895f6f55/CONTRIBUTING.md), [2](https://github.com/wild-linker/wild/pull/1653#issuecomment-4017584459), [3](https://github.com/wild-linker/wild/pull/912)) | [](#request-for-help) |
| [rsync](https://rsync.samba.org/) | version ≤3.4.1 | [](#permissive-ai-policy) ([1](https://github.com/RsyncProject/rsync/commit/aa142f08ef31d3ffa8d6b3b8af16d00324a98c1b), [2](https://github.com/RsyncProject/rsync/commit/b905ab23af2d71363271e99e446e8fe0bfc77f7f)) | version ≤3.4.1, `scp` or a FTP client |
| [npmx](https://npmx.dev/) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/npmx-dev/npmx.dev/blob/e3b575ad0760ce79a278e34fc3553020d093e940/CONTRIBUTING.md#using-ai), [2](https://github.com/npmx-dev/npmx.dev/blob/7f2fc1ae716605c87612b0e138aa95a277d44559/.github/copilot-instructions.md), [3](https://github.com/npmx-dev/npmx.dev/pull/1513/changes/557db327a06ffa2d42e67a8df3ccc107381fe063))<br />[](#ai-code-reviews) ([1](https://github.com/npmx-dev/npmx.dev/pull/2183#issuecomment-4104059314)) | [npmjs](https://www.npmjs.com/) |
> A lot of Make implementations exist that are very fit to replace other task runners.
> In fact so many were produced, that one can just look up "GNU Make alternative" and find >10 results to pick the one they like the best. Also see: [wikipedia/Make](https://en.wikipedia.org/wiki/Make_(software))
> Suggestions for alternative programming languages are necessarily more complex than many other categories considered here, as that's one of the core technical choices to be made when starting a new software development project.
> Depending on your needs and technical concerns, you may find any number of different alternative languages useful. For example, when considering alternatives to .NET, Java may be an interesting candidate from the perspective of being a similar VM-based framework.
> * [Zig] - general-purpose, compiled, system programming language: [Strict No LLM / No AI Policy](https://codeberg.org/ziglang/zig#strict-no-llm-no-ai-policy)
> * Other forks of VIm prior to March 20th 2024, build [v9.1.0190](https://github.com/vim/vim/releases/tag/v9.1.0190) or commit [`8950bf7f8b85c1287d4e696965d88091fcc60594`](https://github.com/vim/vim/commit/8950bf7f8b85c1287d4e696965d88091fcc60594)
| Name | Last Untainted Version or Commit ID | Tags and Evidence | Alternative(s) |
|---|:---:|---|---|
| [espeak-ng](https://github.com/espeak-ng/espeak-ng) | [](#request-for-help)| [](#ai-code-reviews) ([1](https://github.com/espeak-ng/espeak-ng/pull/2328#pullrequestreview-3559773058), [2](https://github.com/espeak-ng/espeak-ng/pull/2302)) <br /> [](#permissive-ai-policy) ([1](https://github.com/espeak-ng/espeak-ng/commit/a17b335f9f794f9176bd702ceacb09f34bcb1e16), [2](https://github.com/espeak-ng/espeak-ng/commit/00c9e543c7390c228266de5dcac12e5fb2ca98f9)) | [espeak](https://espeak.sourceforge.net/) (note: heavily outdated and very bad from a modern standpoint), [flite](http://cmuflite.org/) |
| [Nametag](https://github.com/mattogodoy/nametag) | [](#request-for-help)| [](#permissive-ai-policy) ([1](https://github.com/mattogodoy/nametag/pull/126), [2](https://github.com/mattogodoy/nametag/pull/119), [3](https://github.com/mattogodoy/nametag/pull/70)) | [Monica](https://github.com/monicahq/monica) |
| Name | Last Untainted Version or Commit ID | Tags and Evidence | Alternative(s) |
| --- |:---:| --- | --- |
| [ArkType](https://arktype.io/) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://arktype.io/llms.txt), [2](https://github.com/arktypeio/arktype/blob/main/.cursor/commands/armstrong.md), [3](https://github.com/arktypeio/arktype/pull/1553#issuecomment-3672923281)) <br />[](#ai-code-reviews) ([1](https://github.com/arktypeio/arktype/pull/1594)) | [TypeBox](https://github.com/sinclairzx81/typebox) (in grey area) |
| [Electron](https://electronjs.org) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/electron/electron/blob/f4c4cd14ac8eebf4ed33d75ca4f9b35b0e013208/CLAUDE.md), [1](https://github.com/electron/electron/commit/3295d0d4b05bf338427e5f98374dda206b83651f), [2](https://github.com/electron/electron/commit/816e5964fb574585840ec82f7b1e3e99b3f93785)) | Using native GUIs, such as [GTK](https://gtk.org), instead of making a web site |
| [nvm](https://github.com/nvm-sh/nvm) | before [July 2025](https://github.com/nvm-sh/nvm/pull/3609)? [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/nvm-sh/nvm/pull/3609), [2](https://github.com/nvm-sh/nvm/blob/master/AGENTS.md)) | [](#request-for-help) |
| [Valibot](https://valibot.dev/) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/open-circle/valibot/tree/main/prompts), [2](https://valibot.dev/guides/llms-txt/)) <br /> [](https://codeberg.org/small-hack/open-slopware#ai-in-issue-tracker) ([1](https://github.com/open-circle/valibot/issues/1389)) <br /> [](#ai-code-reviews) ([1](https://github.com/open-circle/valibot/pull/1388)) | [TypeBox](https://github.com/sinclairzx81/typebox) (in grey area) |
| [Zod](https://zod.dev/) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/colinhacks/zod/blob/main/AGENTS.md), [2](https://github.com/colinhacks/zod/blob/main/CLAUDE.md), [3](https://github.com/colinhacks/zod/blob/main/.cursorrules)) <br /> [](#sponsored-by-ai) ([1](https://github.com/colinhacks/zod/blob/c7805073fef5b6b8857307c3d4b3597a70613bc2/packages/zod/README.md?plain=1#L40)) | [TypeBox](https://github.com/sinclairzx81/typebox) (in grey area) |
| [Requests](https://github.com/psf/requests) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/psf/requests/commit/b2a1d33f571518ca9a6148e7da787cc5827f897a)) | [`urllib.request` in Python Standard Library](https://docs.python.org/3/library/urllib.request.html) |
| [`c2rust`](https://github.com/immunant/c2rust) | | [Statement from developer](https://github.com/immunant/c2rust/issues/1653#issuecomment-4035773395)<br />[`postprocess`](https://github.com/immunant/c2rust/tree/master/c2rust-postprocess) component uses LLMs<br/>[PR](https://github.com/immunant/c2rust/pull/1614) made with AI
| [`facet`](https://github.com/facet-rs/facet) | Use more specific deserialization libraries and avoid macros that slow things down anyway. [`serde`](https://serde.rs) itself is [tainted by its Palantir-defending author](https://archive.ph/fe9g5). [`nanoserde`](https://github.com/not-fl3/nanoserde) could be suitable for certain use cases. | [Core developer openly uses LLMs for library dev](https://archive.ph/YhcOh) | |
| [`ratatui`](https://github.com/ratatui/ratatui) | [`iocraft`] (<=0.7.17) and [`cursive`]. [`console`] and its family of libraries may also be enough for some use cases. | [Core developer has stated their use of AI for rust code](https://github.com/ratatui/ratatui/discussions/2201) | |
| [`iocraft`] | An older version (<=0.7.17, prior to [Feb 12, 2026](https://github.com/ccbrown/iocraft/commit/1daff8bdef6d6c0f0ceeb89d0308b1990a5fa428)); [`cursive`] and [`console`] | Contains [commits from Claude](https://github.com/ccbrown/iocraft/commits?author=domenkozar) | |
| [rust-analyzer](https://github.com/rust-lang/rust-analyzer) | cargo check | [Allows AI contributions](https://github.com/rust-lang/rust-analyzer/pull/21314/changes) and [has Claude rules](https://github.com/rust-lang/rust-analyzer/blob/master/CLAUDE.md) | |
| [`wgpu`](https://github.com/gfx-rs/wgpu) | [`vulcano`](https://github.com/vulkano-rs/vulkano), [`ash`](https://github.com/ash-rs/ash), [`glow`](https://github.com/grovesnl/glow) | Explicitly [allowed in CONTRIBUTING.md](https://github.com/gfx-rs/wgpu?tab=contributing-ov-file#llms-ai), [LLM instructions](https://github.com/gfx-rs/wgpu/commit/bed71efe59e2360c625163c28c70de598dee41b7) added according to a decision by the maintainers in [this issue](https://github.com/gfx-rs/wgpu/issues/8834) |
| [`zbus`](https://github.com/z-galaxy/zbus) | | [CLAUDE.md](https://github.com/z-galaxy/zbus/blob/78f786cad319f9027a9893f73aa862ed2f4b45cd/CLAUDE.md), [core developer talking about their AI policy](https://matrix.to/#/!uSaWOSkfhbBXoCCxYe:matrix.org/$6frxoLIKKBIE7ejJr3OJk9jn_nSTl_B5a2poK4lxDjo?via=matrix.org&via=gnome.org&via=mozilla.org) ([screenshot](./evidence/img/zbus.png)) |
| [UPBGE 0.5](https://upbge.org) | UPBGE 0.36.1 seems to be slop-free | [0.5 release notes state the use of AI in plain language](https://github.com/UPBGE/upbge/wiki/Release-notes-version-0.50#b-implementation-technique-and-limitations) |
| [macports](https://github.com/macports/macports-ports) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/macports/macports-ports/pull/28628)) | Build packages from source |
| [PiKISS](https://github.com/jmcerrejon/PiKISS) | [](#request-for-help) | [](#ai-code-reviews) ([1](https://github.com/jmcerrejon/PiKISS/pull/240)) | Use system repositories provided by your distro or manual installation of software |
> * [KeePass 2](https://keepass.info/download.html). It is a .NET application, but [Mono](https://keepass.info/help/v2/setup.html#mono) or [Wine](https://keepass.info/help/v2/setup.html#wine) can be used for non-Windows platforms.
> * [KeePassXC 2.7.9](https://github.com/keepassxreboot/keepassxc/releases/tag/2.7.9) was released before the statement, and was awarded 3 year security Visa by the French National Cybersecurity Agency ([ANSSI](https://cyber.gouv.fr/)) for a First-level Security Certification (CSPN), valid in France and Germany, under report No. ANSSI-CSPN-2025/16 ([archive.org](https://web.archive.org/web/20251128131437/https://cyber.gouv.fr/produits-certifies/keepassxc-version-279), [certificate](https://messervices.cyber.gouv.fr/visas/ANSSI-CSPN-2025-16-certificat.pdf)).
> * [KeePassChi](https://codeberg.org/keepasschi/keepasschi): A fork of KeePass 2.7.9 [with clear stance against LLMs](https://social.anoxinon.de/@whitequark@treehouse.systems/116403781772180051).
> * [Jekyll](https://jekyllrb.com/) maybe? No signs of AI in the repo but also [no clear statement](https://talk.jekyllrb.com/t/ai-usage-in-the-jekyll-project/10269) provided
| Name | Last Untainted Version or Commit ID | Tags and Evidence | Alternative(s) |
|---|:---:|---|---|
| [Calibre](https://calibre-ebook.com) | [`8.15.0`](https://github.com/kovidgoyal/calibre/releases/tag/v8.15.0) | [](#ai-functionality) ([1](https://github.com/kovidgoyal/calibre/blob/master/Changelog.txt)) | [BookLore](https://booklore.org/) <br /> [Clbre](https://github.com/grimthorpe/clbre), a Calibre fork without AI <br /> [Arcalibre](https://codeberg.org/rereading/arcalibre), a Calibre hard-fork from before AI features |
| [somafm_tui](https://github.com/zsh-ncursed/somafm_tui) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/zsh-ncursed/somafm_tui/commit/bd176678d3f6b4f63603c331fb4481356068e78e), [2](https://github.com/zsh-ncursed/somafm_tui/commit/11c9e9e934aa0d6fe4a1f572eed2ba252aac064b)) | [soma-player](https://github.com/mpuccini/soma-play) |
| [Nextcloud](https://nextcloud.com/) | [seafile](https://github.com/haiwen/seafile) for file storage and syncing only (not a complete worksuite replacement) | [✨ Nextcloud Assistant](https://nextcloud.com/assistant/) (can be disabled)<br />Nextcloud Desktop has [AGENTS.md](https://github.com/nextcloud/desktop/blob/master/AGENTS.md) |
| [SeaweedFS](https://seaweedfs.com/) | [Garage] for S3 api + [JuiceFS](https://juicefs.com/en/) for the NFS mounts and K8s CSI driver<br />[Rook/Ceph](https://rook.io/) | Using [Gemini and coderabbitai in PRs](https://github.com/seaweedfs/seaweedfs/pull/7996). |
| [Mastodon](https://joinmastodon.org/) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/mastodon/.github/blob/49b9c64f3e2b6a459e56a05068c05b3247659bb3/AI_POLICY.md)) | See below |
| [Misskey](https://misskey-hub.net/) | [](#request-for-help) | [](#vibecoded) ([1](https://github.com/misskey-dev/misskey/commit/2fa6ecc7efaaf9b9d189cdd3a3ebbb9171c86078)) | See below |
| [ntfy-sh](https://ntfy.sh/) | see below | [v2.18.0 is "14,997 added lines of code \[...\] written by Cursor and Claude"](https://github.com/binwiederhier/ntfy/releases/tag/v2.18.0) ([archive](https://web.archive.org/web/20260308142510/https://github.com/binwiederhier/ntfy/releases/tag/v2.18.0)) |
list of unified push distributors at: https://unifiedpush.org/users/distributors:
- Sunup
- NextPush (see entry for NextCloud on taintedness)
| [lvm2](https://sourceware.org/lvm2/) |[](#request-for-help) |[](#permissive-ai-policy) ([1](https://gitlab.com/lvmteam/lvm2/-/commit/6207fe707a4ae255ef62a2fa088ec2497ae6c0a8),[2](https://gitlab.com/lvmteam/lvm2/-/commit/ff76548a3da5cee224e4663d81d7041558115b8e),[3](https://gitlab.com/lvmteam/lvm2/-/commit/99b85e7e4acc7fb69460b2bdb94aa8c2ca0f8946)) | [](#request-for-help) |
| [rsyslog](https://www.rsyslog.com/) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://www.rsyslog.com/doc/about/ai_first.html)) | [syslog-ng](https://www.syslog-ng.com/products/open-source-log-management/) |
| [systemd](https://github.com/systemd/systemd) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/systemd/systemd/blob/main/AGENTS.md), [2](https://github.com/systemd/systemd/commit/744d589632c545e90ae76853abbfbc90cb530e24)) | [dinit](https://davmac.org/projects/dinit), [s6](https://skarnet.org/software/s6/), [OpenRC](https://github.com/OpenRC/openrc), [shepherd](https://shepherding.services/) |
| [wireplumber](https://gitlab.freedesktop.org/pipewire/wireplumber/) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://gitlab.freedesktop.org/pipewire/wireplumber/-/blob/master/AGENTS.md?ref_type=heads)) | [](#request-for-help)|
| [Mesa3D](https://mesa3d.org/) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://docs.mesa3d.org/submittingpatches.html#expectations-on-contributors)) | [](#request-for-help)|
| [Limine](https://github.com/Limine-Bootloader/Limine) | [Hyper](https://github.com/UltraOS/Hyper), [GRUB](https://www.gnu.org/software/grub/) | This entry was added by the creator and main maintainer of Limine. Limine does not ban code authored by LLMs or anyone/anything else, as long as it's reviewed by a human. The author has used LLMs (Claude Code) numerous times for many commits. |
| [FreeBSD](https://freebsd.org) | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://github.com/freebsd/freebsd-src?tab=contributing-ov-file#quality-expectations)) | See below |
| [GNU Mach](https://www.gnu.org/software/hurd/microkernel/mach/gnumach.html) - kernel for GNU Hurd | [](#request-for-help) | [](#permissive-ai-policy) ([1](https://lists.gnu.org/archive/html/bug-hurd/2026-02/msg00133.html)) | See below |
> Operating systems are a tough ask. No matter what you choose, support will be inferior compared to an operating system with the Linux kernel, so don't feel pressured to switch.
> * [MirBSD](https://mbsd.evolvis.org/) has [banned LLMs](https://mbsd.evolvis.org/permalinks/wlog2021_e20240726.htm#e20240726_wlog2021)
> * [NetBSD](https://www.netbsd.org/) has a [somewhat ambiguous policy](https://www.netbsd.org/developers/commit-guidelines.html)
> * [QEMU]: Emulator and virtualizer that [does not accept slop contributions](https://www.qemu.org/docs/master/devel/code-provenance.html#use-of-ai-generated-content).
> * [virt-manager]: GUI for managing libvirt machines. Needed for VirtualBox feature parity as QEMU [does not provide one of its own](https://wiki.archlinux.org/title/QEMU#Graphical_front-ends_for_QEMU).
> * Use [WireGuard](https://www.wireguard.com/)<sup>®️</sup> capable VPN's such as those listed in the official [wireguard repos](https://www.wireguard.com/repositories/).
LLMs are often trained on, and thus prone to, regurgitate either completely, or in-part, chunks of code that are licensed under terms which have specific legal requirements that a sloperator may not understand or even be aware of when making a contribution. Regardless of this ignorance, it falls to the repo's owner to comply with the terms of any and all licensed code integrated into their project.
* [copilot litigations](https://githubcopilotlitigation.com/), [IEEE article explaining how we go here](https://spectrum.ieee.org/ai-code-generation-ownership)
* [broader lawsuits against AI companies tracker ](https://chatgptiseatingtheworld.com/2025/11/02/tracker-of-tort-lawsuits-v-ai-companies/)
Legal, copyright and ethic problems arise especially with copyleft licenses such as (A/L)GPL. With the "help" of AI the copyleft code may be "license-washed" very easily.
There are ongoing problems with AI "license-washing" in the FOSS world:
AI companies use data from across the web for training their models, most often without the website owners' and users' consent. Big tech companies like Google and Meta are scraping data from the users of major FOSS projects, such as Mastodon, WordPress, and other ActivityPub-powered and self-hosted software.
* In 2023, [the Washington Post published a list of sources in Google's C4 data set](https://archive.ph/eehKq). A multitude of fediverse instances and personal sites were included. The fediverse is known for its userbase being major proponents of privacy and opt-in consent, making this especially jarring for those who have chosen to use decentralized social media for control over their data.
* In 2025, [a similar leak of Meta's sources was published](https://archive.ph/NZlf3). Meta's list demonstrates how their integration of ActivityPub into their Threads software has enhanced their ability to mine content without authorization. Threads is widely blocked in some parts of the fediverse, but their scraping of server CDNs has allowed them to get around that. Notably, both the CDN domains of the managed hosting services masto.host and fedi.monster are included in the list; large servers like mastodon.art, which is hosted by the former and has many artists who've left sites like DeviantArt and others due to their AI scraping of user content, had [media unknowingly scraped](https://mastodon.art/@Curator/115022115346692178).
FOSS projects listed in this repo are using tooling that blatantly disregard licensing and violate of Codes of Conduct, making said tools antithetical to FOSS' purpose.
To start learning a bit more, you can checkout the wikipedia page on [Environmental impact of artificial intelligence](https://en.wikipedia.org/wiki/Environmental_impact_of_artificial_intelligence#Individual_level). We're very open to people contributing other explanations, links, and resources to learn more about this. Here's what we've gathered so far:
- [MIT Technology Review: We did the math on AI’s energy footprint. Here’s the story you haven’t heard.](https://www.technologyreview.com/2025/05/20/1116327/ai-energy-usage-climate-footprint-big-tech/)
AI usage and normalization contributes to labor violations in many ways that are obvious and some you may not be aware of.
On one hand, many things that you think are "AI" are actually humans in another country pretending to be an AI chatbot for you for either extremely low wages or in some cases, no wages e.g. prison labor. This is particularly common for "friend"/"sex" bots, but it is also extremely common in the image/video identification. You can find a bit more info at the following links:
- [Long hours and low wages: the human labour powering AI’s development](https://theconversation.com/long-hours-and-low-wages-the-human-labour-powering-ais-development-217038)
- [OpenAI Used Kenyan Workers on Less Than $2 Per Hour to Make ChatGPT Less Toxic](https://time.com/6247678/openai-chatgpt-kenya-workers/)
- ['AI Is African Intelligence': The Workers Who Train AI Are Fighting Back](https://www.404media.co/ai-is-african-intelligence-the-workers-who-train-ai-are-fighting-back/)
- [These Prisoners Are Training AI: In high-wage Finland, where clickworkers are rare, one company has discovered a novel labor force—prisoners.](https://web.archive.org/web/20260108224812/https://www.wired.com/story/prisoners-training-ai-finland/)
- [Amazon grocery stores previously reported to use AI actually used people in India](https://www.businessinsider.com/amazons-just-walk-out-actually-1-000-people-in-india-2024-4?op=1)
Vibe coding / agentic workflows result in poorer code quality, and relaxed oversight practices. These effects may be compounded by the common practice of using additional LLM-based tooling to provide code-reviews.
* [How I Dropped Our Production Database and Now Pay 10% More for AWS](http://archive.today/2026.03.06-144058/https://alexeyondata.substack.com/p/how-i-dropped-our-production-database)
* [Claude Tested Everything Except the One Thing That Mattered (Ai agent refuses to follow explicit instructions to test `createPost()` in increasingly erratic ways)](http://archive.today/2026.03.09-201135/https://christophermeiklejohn.com/ai/claude/2026/03/08/claude-tested-everything-except-the-one-thing-that-mattered.html)
* [Amazon calls engineers for a “deep dive” internal meeting to discuss “GenAI”-related outages](https://ghostarchive.org/archive/3TfgF)
* GitClear has released reports in [2024](https://www.gitclear.com/coding_on_copilot_data_shows_ais_downward_pressure_on_code_quality) and [2025](https://www.gitclear.com/ai_assistant_code_quality_2025_research) indicating a worsening of key code quality metrics correlating with increased LLM adoption.
There is increasing evidence to show that LLMs negatively impact developers' coding abilities:
* [Brains show less activity when completing tasks with LLMs](https://arxiv.org/abs/2506.08872) compared to completing tasks with search or completing tasks without digital help.
* [Developers who use early-2025 LLMs reported higher subjective performance](https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/), but were measured to have lower objective performance. This gap between subjective and objective performance was considered notable.
* [In an Anthropic study](https://www.anthropic.com/research/AI-assistance-coding-skills), learners using LLMs demonstrated lower learning rates on average compared to learners not using LLMs.
* [A recent study uses the term "cognitive surrender"](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6097646) to describe the way humans tend to offload key critical thinking skills onto LLMs, even when the output is wrong.
* There are [anecdotal reports of LLM users forgetting how to code](https://www.theverge.com/ai-artificial-intelligence/767973/vibe-coding-ai-future-end-evolution).
* [Meta Security Researcher's AI Agent Accidentally Deleted Her Emails](http://archive.today/2026.02.26-153034/https://www.pcmag.com/news/meta-security-researchers-openclaw-ai-agent-accidentally-deleted-her-emails)
* [Moltbook’s ‘vibe-coded’ breach is the future of security failures](https://thehill.com/opinion/cybersecurity/5744310-ai-powered-security-risks/)
* [In a study evaluating over 500k code samples, LLM-generated code was found to contain more high-risk security vulnerabilities than human-generated code](https://arxiv.org/abs/2508.21634)
* [LLMs make up package names, making them vulnerable to incorporating malicious code in "slopsquatting" attacks](https://www.theregister.com/2025/04/12/ai_code_suggestions_sabotage_supply_chain/) ([Arxiv study](https://arxiv.org/abs/2406.10279))
* [Wikipedia: Deaths linked to chatbots](https://en.wikipedia.org/wiki/Deaths_linked_to_chatbots)
* [AI chatbot pushed a teen to kill himself, a lawsuit against its creator alleges](https://apnews.com/article/chatbot-ai-lawsuit-suicide-teen-artificial-intelligence-9d48adc572100822fdbc3c90d1456bd0)
A lot of AI companies also work directly with nation states for use in their Departments of War (sometimes called Defense) which in turn leads to further AI usage during war and invasions. This is coupled with [NYT: Palantir, Anthropic and small start-ups are generating rewards from their investments in defense tech](https://www.nytimes.com/2026/03/18/technology/silicon-valley-war-defense-tech.html).
As another example [NPR: OpenAI announced Pentagon deal after Trump banned Anthropic](https://www.npr.org/2026/02/27/nx-s1-5729118/trump-anthropic-pentagon-openai-ai-weapons-ban) which was due to the USA Department of War [launching an AI acceleration strategy](https://web.archive.org/web/20260113071131/https://www.war.gov/News/Releases/Release/Article/4376420/war-department-launches-ai-acceleration-strategy-to-secure-american-military-ai/).
Due to the nature of LLMs being only kind of as good as the data they are trained on, this can lead to additional civilian deaths and housing/infrastructure damage either intensionally or not. Examples:
- [Gaza: UN experts deplore use of purported AI to commit ‘domicide’ in Gaza, call for reparative approach to rebuilding](https://www.ohchr.org/en/press-releases/2024/04/gaza-un-experts-deplore-use-purported-ai-commit-domicide-gaza-call)
- [Lavender & Where’s Daddy: How Israel Used AI to Form Kill Lists & Bomb Palestinians in Their Homes](https://www.democracynow.org/2024/4/5/israel_ai)
- [Microsoft says it provided AI to Israeli military for war](https://apnews.com/article/microsoft-israel-military-gaza-hamas-artificial-intelligence-20b2adb438b39ee9cb6eb2f52c1ae44a)
- [Google has dropped its promise not to use AI for weapons](https://theconversation.com/google-has-dropped-its-promise-not-to-use-ai-for-weapons-its-part-of-a-troubling-trend-249169)
</details>
All of this to remind you that if you use AI, you're helping to support these companies and the additional activities they participate in, outside of generative code or images.
Police have quickly embraced AI, which has already directly led to people being jailed for things they've never done. As examples:
- [This Grandmother was jailed for 6 months after an AI error linked her to a crime in a state she had never even visited](https://www.the-independent.com/news/world/americas/crime/tennessee-grandmother-ai-arrest-error-north-dakota-b2938261.html)
- [How Wrongful Arrests Based on AI Derailed 3 Men's Lives](https://web.archive.org/web/20260111202754/https://www.wired.com/story/wrongful-arrests-ai-derailed-3-mens-lives/)
This is, in part, due to companies such as Amazon [Aggressively pushing police to use AI](https://web.archive.org/web/20260116082026/https://www.forbes.com/sites/thomasbrewster/2025/10/01/inside-amazons-aggressive-push-to-get-cops-using-ai-surveillance/) which they do through both facial recognition and offering compute for predictive policing. With regards to facial recognition, here's an example of how it too can lead to false arrests: [Face Recognition on Flawed Data](https://www.flawedfacedata.com/#art-or-science).
- [Bennan Center of Justice: The Dangers of Unregulated AI in Policing](https://www.brennancenter.org/our-work/research-reports/dangers-unregulated-ai-policing)
- [OxJournal: Predictive Policing or Predictive Prejudice? A Study of the Legal, Historical and Ethical Implications of AI in Policing](https://www.oxjournal.org/predictive-policing-or-predictive-prejudice/)
- [The Guardian: Police AI Chief admits crime fighting tech will have bias...](https://web.archive.org/web/20260224070735/https://www.theguardian.com/technology/2026/feb/24/police-ai-chief-admits-crime-fighting-tech-will-have-bias-but-vows-to-tackle-it)
- [A third of all Black children were flagged by a child services agency](https://loganstapleton.com/wp-content/uploads/2022/04/Extended_Analysis__How_Child_Welfare_Workers_Reduced_Racial_Disparities_in_Algorithmic_Decisions.pdf)
Having to deal with the onslaught of many LLM written pull requests and issues, causes real maintainer burnout that stagnates projects as maintainers become overwhelmed with half baked, poorly written, insecure code. Here's some examples:
* [Overrun with AI slop, cURL scraps bug bounties to ensure “intact mental health”](https://arstechnica.com/security/2026/01/overrun-with-ai-slop-curl-scraps-bug-bounties-to-ensure-intact-mental-health/)
* [Godot maintainers struggle with 'draining and demoralizing' AI slop submissions](https://www.theregister.com/2026/02/18/godot_maintainers_struggle_with_draining/)
* [An AI Agent Published a Hit Piece on Me (matplotlib)](http://archive.today/2026.02.19-045842/https://theshamblog.com/an-ai-agent-published-a-hit-piece-on-me/)
The demand for construction and outfitting of new data-centers to host AI/LLM compute capacity has overwhelmed global supply and production of multiple hardware components. Memory, Storage, and GPUs have seen massive increases in price for both consumer and enterprise models upward of 400% in some cases.
The lack of supply has led large system-builders to purchase production capacity from OEMs well in-advance of delivery leading some manufacturers to end consumer-oriented product lines in favor of enterprise capacity.
The down-stream effects for consumers is that near all electronic devices which contain memory and storage will see their prices rise and availability decline. Those who already own existing electronics and computer hardware components may also find themselves unable to have their devices repaired or replaced under warranty.
- [The RAM shortage is coming for everything you care about](https://www.theverge.com/tech/880812/ramageddon-ram-shortage-memory-crisis-price-2026-phones-laptops)
- [Western Digital is already sold out of hard drives for all of 2026 — chief says some long-term agreements for 2027 and 2028 already in place](https://www.tomshardware.com/pc-components/hdds/western-digital-is-already-sold-out-of-hard-drives-for-all-of-2026-chief-says-some-long-term-agreements-for-2027-and-2028-already-in-place)
- [The 2026 storage crisis: Why AI data centers are hoarding every hard drive on the market](https://www.howtogeek.com/dont-count-on-hdds-to-save-you-from-rising-storage-costs/)
This all results in shrinking the pool of people who have access to building home computers for any purpose, from gaming to coding to home labs, which in turn makes the tech industry less diverse due to people who have been historically marginalized having less financial resources to learn the skills at home. When this is factored in with the price of college being unaffordable in many places, we will see a sharper decline in disabled people, people of color, women, and the queer community entering the tech industry.
This repository is licensed under the Creative Commons Attribution Share Alike 4.0 International license. Please see [LICENSE.txt](LICENSE.txt) for more information.
