Interview with Katharina Meyer: On the tension between public interest and profit maximisation in public interest tech

This is an interview with Katharina Meyer, who has been a member of several technology funds in Germany in recent years. The interview was conducted in February 2024. In this interview she describes her thinking in the role of a funder and enabler of public interest technology, around how they could decide if a project is in the public interest or constitutes a public good, and various interactions between such projects/products and the market, including the importance of free and open source solutions.

Katharina Meyer is a historian of technology and science. She was Head of Research for Public Interest Technology at the Prototype Fund from 2018 to 2021, a project of the Open Knowledge Foundation and the German Federal Ministry of Education and Research (BMBF). and led a project on Data Commons at The New Institute in Hamburg. Subsequently she headed the Research in the development of the Sovereign Tech Fund. Since April last year she is the Director of the Digital Infrastructure Insights Fund, a global initiative that provides a platform and funding for academics and practitioners to better understand how open digital infrastructure is built and deployed.

Theresa Züger: In your work you get very deep insights regarding digital infrastructures that are crucial for the digital world as we know it, and infrastructure that could also be defined as public interest infrastructures. How do you like this terminology? Do you like this description? And what is public interest tech for you in general?

Katharina Meyer: I liked the terminology. When we conceptualised what the Digital Infrastructure Insights Fund as a research entity is set up to do, we also had to talk about terminologies a lot and think about the fairest way we could approach the public interest as a concept. And our understanding back then was that we define public interest infrastructures as: all the open digital infrastructures and the corresponding knowledge assets plus norms and standards that form the base layer of the supply chain for software and connectivity, – used by governments, by civil society, and businesses alike – often produced by volunteer communities and subject to market failures. We tried to deploy the quadruple helix as another defining factor for who we intended to reach with our resources and insights. The public interest is still hard to define because you have to specify in each sector what it entails, which challenges the specific sector faces: is it market consolidation, or is it something else that needs to be addressed? So in general, I like public interest as an aspiration, but it's also challenging in regard to terminology. It is challenging to determine who are the benefactors and beneficiaries, who are the ones that are exploiting the public interest and why. And especially looking at the constellation between these actors: The governments, the public, businesses, and academia. A narrative needs to be formed. There needs to be a distinct set of measures that you need to define to make informed investments and formulate research hypotheses that will improve the ecosystem for the long-term.

TZ: You mentioned the quadruple helix. What is this?

Katharina Meyer: The quadruple helix stems from innovation theory. Originally it's just a term that describes how academia, government administrations, industry, and civil society including citizens interact when new technologies (or other inventions) are developed. What connections exist between them, what imbalances might exist. In innovation theory, it is used mostly to explain why everything is going well, open innovation is great, and nothing needs to be changed. For our case of public interest tech and the discontents in the social development environments in Free and Open Source Software (FOSS) communities through their exposure to markets: I do believe change and critique are necessary, and the helix as a structure can be helpful here to instruct an overview of actors.

TZ: So what would AI systems require, in your view, to be a truly public interest technology?

Katharina Meyer: In my program we have a few projects exploring the legal context of AI focussing on the policy lens that needs to be applied to all of these different instances and also the regulatory environment: In my opinion, AI is one of the examples where you have to include the entire stack in your equation. Next to hardware (computing power is key), you have to include the data infrastructure and the software that is used to train models or the software that AI is deployed in and meets the customer or citizen. It is one of the all-encompassing applications of the public interest idea. For this reason, it is complex to define the public interest. It involves multiple elements and layers, from industry policy to customer protection regulation and many factors that could make it public interest or on the opposite could prove harmful.

TZ: A study published by the Forum for Information and Democracy demanded that AI become a public good. Do you agree to this demand and what would need to happen and change to make this possible?

Katharina Meyer: If it's a public good then, by definition, the provision of the public good is organised as cooperation between a public funder and sometimes a market that would provide the good under “normal” economic circumstances. So, I think it is problematic to call for the provision of the good by government entities alone (although I’m unsure what AI signifies in this specific report because I haven’t read it). Similarly, social media platforms for communication should be provided by the private or non-profit sector, with the public sector playing a regulatory role. To be honest, I think the state shouldn't be the one providing the technical infrastructure or be the content host for anything, but they should provide the resources to build in the public interest, and the legal framework. There's a fine line there. In regards to public goods, I have no issue with the state enabling health insurance or anything comparable service, which is also a market or club good and wouldn’t otherwise be accessible for the public in the desired quality and for a price that is acceptable. But when it comes to infrastructure that people use to deliberate on, or that has anything to do with the context of a democracy and how it's enacted, I think it can be problematic to resort to just determine something to a be public good in a strict sense of the definition, because then the government might become involved in a way that is not beneficial in the end.

If we are in the realm of data trusts and provisions for the training models of AI applications, or when it otherwise makes sense that the government or a state entity provides guardrails or could act as the fiduciary of the technology, I see no issues with that. But if we get to a point where a government starts to review every application of matter before it gets to the public, I think that is not how technology should be regulated at all.

But: a public good can be detected by market failure. Especially with the recent news about the French company Mistral and how it tried to shield Microsoft's competitors from regulation and now uses the exemptions for its own gain: The fact that it's now a trust/competition case means there´s more market consolidation now. If you look at that, at the distribution of actors: a wide variety of actors and no bottleneck is what we need to go for and not the opposite. And support structures need to factor this in: The more monopolistic the ecosystem becomes, the easier it is to take it down, regarding software security or in regards to political attack on the means of production and technology. So in general, there could be benefits to reclaim a technology ecosystem, or certain components as public good- but it doesn’t work for every aspect of the stack, and you have to be careful with the countermeasures you select.

TZ: There is an ongoing debate on if and how public interest systems can be part of a business model or be part of an organisation that has a business model that follows market incentives and aims for profit maximisation. How do you see this from your experience? How do you see this tension?

Katharina Meyer: I mostly work on or fund research on open source software projects at the infrastructure layer and maintenance questions related to those. And I see there are different types of projects in the open source world. Some of them do implement a business model. And there is nothing wrong with that because they provide services that are of quality and can also be important in business contexts where there are resources available to provide a compensation for your time and service. Then there is other open source infrastructure or application software that cannot have a business model, because it’s not unique enough, or just opaque, or just the bolts in the system – it is “boring” technology at first glance. There will be no revenue model other than taxpayers money and subsidies to make sure that these bolts still exist, and that somebody checks their state from time to time so they don’t take down entire systems. I think both can exist next to each other. If we approach the tensions from the public goods debate, the market itself will not provide a service or good at a quality-level that is sufficient to base your services on. And then public entities can by legal definition become active with regulation and provide additional resources and intervene in different ways.

Financialisation of open source software is not a general concern if governance models are clearly established, if the projects follow the ideals of transparency, immutably open in their core, but adaptable to different causes. But: if profit maximisation is the ultimate goal, I think that introduces a lot of pitfalls because the projects now might have competing rationales.

So, it is a case by case decision if it is an issue or not, same as with the public interest technology definition question.

TZ: I think you made a really interesting distinction. You made this distinction between a business model that is part of the market, but not going for profit maximisation, and other models that are definitely going for profit maximisation. Do you see this difference in a different type of organisation? So, different business models that are also formally manifested or is this also a difference simply in each organisation and how it works? Have you seen examples of how people handle this differently? My worry is that in a market oriented business model, there's always the danger of, if you're successful, aiming for profit maximisation and throwing your initial public interest goals overboard. Do you see this danger and does the difference need to be formal or how do you make this distinction work?

Katharina Meyer: Of course, there are governance models that are probably more beneficial to the public interest. For instance, cooperatives are interesting models in the market. They provide services, they get paid. But growth and profit maximisation is not the ultimate goal. And there are many different value infrastructures, under different legal models that allow you to monetise but invest everything that you earn back into the company or invest in a greater social goal. But the normal example of businesses as we see it nowadays, most of them have profit maximisation logic at heart because it's capitalism we are living under.

There are also other examples that are formalised entities and progressive – like infrastructure provider associations. And for them it seems to work well because they also serve a population and customers that are aligned with their values.
Another interesting option to introduce, especially if the public sector is involved: a dedicated public interest customer. An example for this could be the Sovereign Tech Fund, which works as a public interest customer. It contracts services from maintainers of key open digital infrastructure that the market doesn't provide for but which are critical to the functioning of the project to the public, or to science etc. This intervention then makes sure that the project they're supporting can scale its (social etc.) infrastructure while it scales its services. Their other clients from industry usually don’t include this assessment of what contractors need to thrive. It’s interesting that the government or a government entity can act like that to further public interest, and it's all there in the legal toolset the government has at hand.

TZ: Interesting. Can you give a good example of the conflict between market logic and public interest for technology?

Katharina Meyer: Yes. Just from my realm of work: if a project is incentivised to only factor in and implement the features that its highly paying customers are requesting. People might experience an exhaustion of their time and find that they do a lot of busy work for the commercial aspects their project supports but don't have the freedom to make their own decisions for the development trajectory of their project anymore or focus on what they would like to prioritise. That is a clear conflict. I think here the market dynamics need to be balanced with other resources to make sure these projects still can self-determine and not get eaten by their success. Because self-determination was the initial driving factor of open source projects, especially in infrastructure. They thrive because the core developers can make their own decisions regarding development and maintenance, not only account for external factors. Scratching their own itch, in a way. And if you strip that away or shift its focus entirely, that’s problematic.

TZ: You just used the term open source projects synonymously with public interest tech projects. What is the relation between the two for you?

Katharina Meyer: I think I would use them synonymously when it comes to infrastructure projects, because they are, to my understanding, the core definition of the added value this open type of technology can deliver in the context of public interest. Infrastructural technology has permissive licences to allow people to use them for whatever they are building, mostly for free, just by making sure that the source is clear. This is how the commercial software market could grow so impressively fast – it copy and pastes, and that makes sense. This of course does not follow the classical narrative in technology that important tech is always a great new invention, shiny, and something that nobody has ever seen before. And in fact public interest tech is often just a set of the most boring public works on the internet. And if we didn’t have that, nothing else would work. But it's also not what you get a lot of recognition for as a coder or community shepherd. It’s often invisible. People only complain if it doesn't function anymore. And so that would be a good example for public interest technology in my opinion, because it is created without much concern for the individual benefit but the general benefit it can provide if multiplied.

TZ: And following your argument, that would mean that every public interest technology in your understanding also needs to be open source?

Katharina Meyer: Yes. There are some exceptions, I understand that because of security policy and other geopolitical, political assessments, not every technology can be open source because we don't live in a perfect world with only rational actors. So, there are limits to my definition, of course, informed by real world politics. But when it comes to public services there should be a public stack for every government service that is based on open source. For instance, the German government spends a lot of money on proprietary software, and I don't think that is fair. I don’t think that is a good use of my and other taxpayer’s money because you have to contract the same services over and over again because you can't adapt the system to a changing environment. And I think if all of that money would have been invested in open source projects that provide the same services, they would be in a way better state now and more usable, because I understand that the arguments are right: open source is often ugly and hard to use sometimes, but that is also the product of twenty years of a biased system. Open source systems didn't have to be that way now, if the sovereignty arguments we hear now would have been as important twenty years ago.

TZ: So what, in your view, would be the best solutions that would need to happen to manage this tension between the market logic and the public interest for a technology such as AI.

Katharina Meyer: I think there is a lot of room for improvement in industry policy, but also for clauses in contracts with government and public service providers. The city of Barcelona, for instance: if you want to provide a service for the government you are mandated to work under an open access or open data policy. This can ensure that the taxpayers money is used to pay for services to be open sourced, for instance, or they need to be interoperable and they cannot be locked in service. Also, I could imagine an entity that manages all of the IP for publicly funded research. For instance, vaccinations that would be in the public interest and need to be distributed outside of market rationales. I mean, there are many ways to structure society in this way – with commoning. It is not very typical yet, but it's also not impossible to imagine it differently. People working in government need to be a bit smarter. They need to keep up with legal innovations. They need to keep up with the public interest debate and also what this leads to because we are finding malfunctions left and right in economics. One textbook example would be external costs for the environment: Environmentally harmful goods are too cheap. Costs caused by industrial pollution are not factored in realistically. It is quite clear that we cannot continue in the way it is now. And that is also why there are a lot of initiatives like the STF and other governments providing money to test out a set of new measures to shift markets towards the public interest. And I like that we live in this time.

TZ: That's a very positive note. I'm happy we got to that. So I'm going to come to my last question. Have you seen successful examples of public interest realised by a for profit organisation and what is generally the most important – strategies or values – to make this successful?

Katharina Meyer: I think Signal is quite a successful example. Signal as a company is a good example of providing public interest technology and being invested in public interest standards. And it is an infrastructure (the underlying protocol) as well as an application. That's why I chose that example.