Skip to content

AI_POLICY.md

Latest commit

 

History

History
45 lines (27 loc) · 4.03 KB

AI_POLICY.md

File metadata and controls

45 lines (27 loc) · 4.03 KB
Raw

AI Usage Policy

The argopy developing team has set some rules for generative AI usage. These rules are not definitive and may be updated in the future.

  • All generative AI usage in any form must be disclosed. If AI was used to generate a significant portion of your contribution (beyond simple autocomplete), we ask that you disclose it within the code or PR description. You must state the tool you used along with a sentence saying that the work was AI-assisted.

  • Pull requests and issues created by AI are forbidden. If AI isn't disclosed but a maintainer suspects its use within a PR or issue creation, the PR will be closed. AI assistance can be used for discussions inside PR or issues, but must always be supervized and reviewed by a human eye before submission.

  • Media. Text and code are the only acceptable AI-generated content, per the other rules in this policy.

These rules apply to all contributors, including maintainers.

Human Accountability

The human contributor is 100% responsible for the contribution.

If you submit a Pull Request that includes AI-generated code, documentation, or comments:

  • You must fully understand the code you submit, and the context in which it is included inside the global project or Argopy.
  • You must be able to explain the "why" behind the implementation during the review process.
  • You are responsible for the long-term maintenance of that code.

The people

Please keep in mind that Argopy is developed and maintained by humans being.

It is for us a fundamental aspect of the project that discussions (in issues or PR), even if happenning on github, are made between humans. Knowing this, it will be considered impolite to approach the Argopy dev community with AI-agent.

Ethics, Ecology & Expertise

For more than 20 years, the Argo program has been a key element in the observation and understanding of climate change and its impact on the ocean. Argo is incremental in monitoring how the human-driven climate change negatively impact the ocean state and marine life (check the IPCC special report on ocean or that brief overview also from the IPCC). So, it is of primary and inherent concern to Argo to limit its environmental footprint.

We like to think that Argopy is part of the Argo ecosystem.

As sustainable software developers, the Argopy team aims to better understand the generative AI’s environmental impact, following our preliminary work in monitoring the carbon footprint of maintaining and developing Argopy. But very little information are available on the generative AI’s environmental impact and "current GAI models are deployed mainly in carbon-intensive regions" (Ding et al, 2025).

With that, 2 things to keep in mind when contributing to the development of Argopy :

  • Ethics & Ecology : Not knowing the - presumably very large - impact of generative AI on CO2 emissions (and use of limited Earth resources), it is important to us that we adopt a conservative and limited approach regarding GAI usage in contributing to Argopy in order to limit its environmental footprint.
  • Expertise : Argopy, as part of a public research infrastructure, has more duty in developing/preserving a technical and scientific expertise for the public good, than participating in a blind race to productivity gain (in the social acceleration sense, an argument echoing the slow science approach at work within Argopy). It is thus also important to us that we adopt a conservative and limited approach regarding GAI usage in contributing to Argopy to preserve a good level of human expertise regarding the produced code and our associated ability to disseminate our expertise to the end-users.