This document governs the use of Artificial Intelligence technologies by unyt.org demokratische freie Netzwerktechnologien e.V. in connection with the development of technical products and services. It shall constitute a binding set of rules for all technical contributions, projects, and products for which members, contributors, and other persons associated with unyt.org e.V. bear responsibility.
For the purposes of this Policy, “Artificial Intelligence” shall be understood in a deliberately broad sense as a term encompassing overlapping technical, societal, and normative meanings. Within the field of computer science, generative AI and Large Language Models (LLMs) in particular currently shape decision-making, development processes, and administrative procedures in unprecedented ways. These developments form the focus of the following rules.
(1) The purpose of AI regulation is to ensure the responsible use of AI technologies. Responsible use of AI technologies minimizes burdens on nature and society and, in particular, ensures that errors caused by AI technologies do not consciously or unconsciously leave the sphere of due diligence of the AI user.
(2) The duty of due diligence applies to all persons contributing to the unyt.org e.V., in particular to persons acting in a reviewing capacity. Due diligence shall be understood as the performance of work independently and free from error to the best of one’s knowledge and conscience. As a general rule, the unsupervised use of generative AI technologies shall not be deemed to satisfy the requirement of acting to the best of one’s conscience.
The use of generative AI technologies is generally prohibited. This applies, for example, to the following purposes:
The use of generative AI technologies is permitted for the following purposes and is subject only to the general requirement of due diligence:
(1a) The use of generative AI technologies shall be disclosed. Such disclosure shall be made as close as possible to the affected portion of the work.
(1b) An exception to this requirement shall apply to the predictive generation of minimal and immediately verifiable code fragments, i.e., “tab completion” and comparable functionality.
(2) Disclosures covering entire contribution areas or collections of contributions, including but not limited to commits and files, shall be inadmissible pursuant to Section 2.
(3) Configurations, references to specific technologies, and instructions relating to the use of generative AI technologies shall be maintained exclusively in a local, i.e., private, manner by the AI user. This requirement also applies to references to AI-specific files in metadata files such as .gitignore.
(1) The use of generative AI technologies should preferably be implemented locally. Where third-party service providers are involved in the provision of AI technologies, such use shall be subject to the Association’s data protection requirements.
(2) Contributions designated as confidential or otherwise subject to special protection of intellectual property shall generally not be processed or utilized using AI technologies provided by third-party service providers. Deviations from this rule shall be permissible only in justified cases and with the approval of the Executive Board of unyt.org e.V.