AI at Prototype Fund?

Like many other funding organizations, at Prototype Fund we are currently looking at how recent developments in AI are affecting our work. As in other industries, the fact remains that - whether we like it or not - the growing use of AI tools is changing our daily lives and the way we approach funding projects. The number of applications is rising not only for us but for most funders: This leads to more intensive selection processes. At the same time, we are confronted with new questions, such as the eligibility of people with less expertise (“vibe coding”) or the limits of using AI tools in funded work.

To address these issues, we’ve established a few simple guidelines for using AI tools at Prototype Fund.

This document is not about whether we fund AI projects, but rather about the use of AI tools within the context of Prototype Fund. Information about our funding focus areas can be found on our application page.

1. No ban on AI in the application process

Prototype Fund supports people in implementing their own ideas. We have no interest in ideas conceived by machines and ask that you refrain from submitting them.

However, if you have a good idea, a clear plan for its implementation (strategy, milestones, etc.), and would like to use AI tools to improve your application, you are welcome to do so.

For nearly ten years, we have strived to make our application processes as accessible as possible. In this context, AI-based translation or text-enhancement tools help break down barriers for people with little writing experience and non-native speakers.

However, this does not mean that we expect all applications to the Prototype Fund to be linguistically polished using AI tools. A submission with a few minor errors is no less likely to receive funding, as long as the text is generally understandable.

2. Vibe coding is not eligible for funding

Unlike writing skills, technical expertise is at the heart of our selection criteria for funding: Prototype Fund specifically funds people who have the necessary software skills to bring their own project ideas to life.

Outsourcing the technical implementation of the project to external parties has always been prohibited; this also applies to outsourcing work to machines.

Please apply only if you (either individually or as th team officially listed in the application) have the necessary programming skills for your project.

To that end, providing a GitHub, GitLab, Codeberg account, or similar will be mandatory starting with applications for the class 03. We want to ensure that applicants are actual developers.

3. AI tools during the funding period? Your decision, your responsibility

As in all areas of Prototype Fund, we give applicants and grantees as much freedom as possible to decide how they want to carry out their work. After all, they are the experts in their own fields, and our role as a funder to decide how grantees should accomplish their work. In short: We fund people who think AI is great, and others who believe it doesn’t belong in FOSS.

As a general rule, we do not rule out the use of AI tools during the funding period. However, our grantees are selected for funding based on their expertise; in other words, a jury chose them (not a machine) because they are confident that they can successfully complete the project.

If you are funded by Prototype Fund, it is up to you to decide what tools you use to this effect. Ultimately, you are the one responsible for the quality of the results - for example, if you are unaware of security vulnerabilities that an AI has built into your software.

In other words: In the context of the Prototype Fund, AI can be used as a tool, but it cannot replace the work of software developers. If code is generated using AI tools, grantees should carefully review the results.

4. Ethical considerations, sustainability, data protection

The environmental sustainability of the selected projects, data protection, and ethical considerations have been priorities within Prototype Fund even before the rise of AI. These priorities have long been an integral part of our selection process and apply equally in the context of AI tools, for example:

• We believe that everyone has a responsibility to care for the environment. For a government-funded funding program like Prototype Fund, this means that we believe that public funds must not be used to develop innovations that harm the climate. When selecting projects, a key question should therefore always be: “Does an idea really need to be implemented in the way described - or can it be done more simply and with fewer resources?”
• Our jury is deeply committed to data protection. Projects that pose a risk to personal data are generally not selected for funding.

And which rules did we set for ourselves?

Our AI guidelines naturally apply not only to applicants and grantees, but also to our work as a team. In short, we have established the following guidelines for ourselves:

• We make sure to use AI tools thoughtfully and sparingly in our day-to-day tasks. These are extremely helpful for certain aspects of our work: For example, they allow us to create all communication materials related to the grant program in two languages, thereby lowering barriers for non-native speakers. To this end, we carefully select our tools and actively ensure that no personal data is shared with third-party companies when using them. We carefully review all results generated by AI before using them further.
• Even though the use of AI in selection processes is currently very much in vogue in the funding world, no AI tools are used in the evaluation of your applications at Prototype Fund. Applications are read and evaluated by people, just as they have been since our very first funding rounds. Any changes to how we deal with applications in regard to the use of AI in the future will be communicated publicly.