2 hours ago
8
Following a wave of genAI code submissions
The foundation behind the open source Godot game engine have announced a crackdown on the use of generative AI by contributors, following a wave of "demoralizing", "low-effort slop" that has added to an already overwhelming quantity of requests for code review. While they'll still let people deploy the chatbot technology for smaller, "menial" jobs such as finding and replacing code, they're taking a stricter stance on "vibe coding" – aka, asking a bot to generate whole swathes of code – and will also require contributors to disclose any genAI usage.
What does this mean for you and I non-technological folk, who still think computers are magic mirrors possessed by unclean spirits? Well, Godot is the engine used by Slay the Spire 2, Buckshot Roulette, The Case of the Golden Idol, Cruelty Squad and Lucid Blocks, among other freaky little projects of note. It's been especially popular since rival Unity torpedoed its own reputation by asking that developers pay up every time somebody installs a Unity game. The nature of its on-going development will impact any number of creators of smaller, stranger games.
Godot runs heavily on the labour of volunteer contributors and seasoned "maintainers". Contributors working on the engine don't simply add their code right to the main Github branch; they submit it for maintainer review in the form of a "pull request". Due to the core team's small size, there are a lot of open pull requests – as the foundation's own management acknowledge in a new blog post, it's become something of a meme. They think it's a sign of health in some ways – it shows that the engine is popular, that they're being "cautious about feature creep", and that they're "dedicated to high code quality". But the problem is largely "the number of qualified reviewers is small", and it's gotten worse since contributors started using generative AI. Some of the newer "contributors" are, in fact, generative AI agents.
"The amount of effort required to make a PR has gone down (and number of PRs has increased as a result), while the amount of work to review PRs and the amount of people available to review has stayed the same," the foundation write. "This reviewer shortage was already a problem, but it was one that we successfully ignored. We can no longer ignore it."
It's not just the quantity of slop that's bothersome; Godot is a community project in which one generation of volunteers trains up the next. The use of generative AI threatens to fracture that community. "AI contributions have the added pain of being demoralizing," the post continues. "Reviewing PRs is already tedious work, but it is rewarding because reviewers generally feel that their efforts are contributing to educating a new contributor (who may become a future maintainer/reviewer). If your feedback on PRs is just being absorbed by a machine and not going towards mentoring a potential future maintainer, it becomes much harder to justify spending your free time on PR review."
With regard to genAI agents submitting code, the Godot Foundation note that large language models "can’t learn from specific feedback and thus can’t benefit from maintainers providing feedback". They add that "AI cannot take responsibility, and we can’t trust heavy users of AI to understand their code enough to fix it."
As such, the Foundation have announced a new, stricter set of polices around generative AI that are designed to train up more actual human contributors as maintainers, and "ensure that people who choose to review PRs feel their time is well spent". Here are the new rules in full:
No autonomous AI agent use or vibe coding- This already leads to an auto-ban from our GitHub repository and will continue to do so.
No use of AI to generate substantial pieces of code
- We require all code to be human authored. AI assistance should be limited to menial things (like code completion, regex, or find and replace).
- If you do use AI in some capacity to author code, you must disclose it in the PR discussion.
No AI-generated text in human-to-human communication
- When our maintainers volunteer their time to review your issue, PR, or proposal, they do not want to talk to a machine. This is a basic principle of respect.
- Machine translations are still acceptable as long as the original content was written by a human.
All PRs must be reviewed and approved by a human before merging - This is the case already, but we will make it more explicit in our policy.
GenAI provisions aside, Godot will now block "new features or significant re-factoring from new contributors without explicit permission from maintainers" to ensure that new contributors – that is, somebody with three or fewer approved pull requests - "take the time to learn the codebase and engage with maintainers to build trust by working on bug fixes and documentation before diving into significant projects".
The Godot Foundation's guarded stance around generative AI is a world apart from the approach taken by Epic with their forthcoming Unreal Engine 6, which will allow developers to "mix and match the best leading-edge models and build custom integrations of all sorts". A few erstwhile Epic partners and Unreal Engine developers have distanced themselves in response. I'm not sure whether Unreal Engine 6's creators made significant use of genAI to make the thing, but going by CEO Tim Sweeney's enthusiasm on the subject - which extends to deriding Epic's competitor Valve for requiring generative AI disclosures on Steam - I'd be surprised if they didn't.
English (US) ·