The GitHub policy paper on the AI Act can be found here.
Hello. Thank you for having me here. Thank you to those watching online. I’m Thomas, and I’m a developer. And you know, I’ve been a developer so long—I normally see a few people in the audience I know. I have been to a LOT of developer conferences, but not to discussions about public policy. I must admit: I probably don’t know most of you.
Which is why I’m so excited to be here today. It’s my first time giving remarks on any public policy, ever. I’m here in Brussels, because the AI Act will define how the world regulates AI. And we need to get it right, for developers and the open source community. For those that produce the software we all depend on.
Now, if you couldn’t tell from my accent, yes, I am German. I grew up in a small apartment in Berlin. In a suburb called Marzahn on the East side of the wall, before it fell. My life has transported me around the world, allowed me to hold positions like the one I hold today, in no small part because I’ve always had just one passion: to build with code.
Now, it is my responsibility as CEO of GitHub to help a new generation of developers, in every way I can, to build the next great thing of tomorrow. And over the course of my life as a developer, I have witnessed so many changes to the nature of software development. Today, I want to focus on two of these shifts: open source and AI.
I came up with the rise of open source. Looking back to the late ‘90s and through the 2000s, the community was met with a lot of skepticism.
People bet against open source. 30 years later, the doubters have been proven resoundingly wrong. The collaboration model open source provides is just better. Today, open source powers the digital world we live in—it powers governments and corporations alike.
And open source is at the heart of European values. It is built with deliberation, open contributions, and diverse perspectives that know no borders. And the ultimate belief that everything can be better with collaboration. It’s because of this that open source is powering a spring of innovations across the world and right here in Europe.
As the home of open source at GitHub, we have long believed it is our responsibility to help protect and empower developers’ rights to innovation and equal opportunity. We’ve been fighting for European developers when it comes to public policy, from the Copyright Directive to the DSA—to today, as the EU considers CE marks for software.
Open source changed the nature of software development and fueled the world’s progress. Now we’re witnessing the second wave of innovation that is revolutionizing how software gets made: the rise of Artificial Intelligence.
We are just beginning the age of AI. It is the defining technology of our times. And it’s already proliferating across all parts of work and life. It’s sorting our photos on our phones. It’s picking our weekly playlist. AI is driving our cars.
AI-powered innovations are sweeping over every industry and every country. Developers are behind all of this, and they’re collaborating on the latest AI innovations–in the open–on GitHub.
And now even developers themselves are increasingly using generative AI to build software, including with GitHub Copilot– a tool that allows them to write their code in half the time. To me, that’s what’s special about this moment: the convergence of AI and open source.
Open source has been at the core of AI development. Open source frameworks like PyTorch and TensorFlow power nearly all of AI, and open source tools have helped improve transparency and remove bias for years. And now AI models are built and shared on open source. The OSS community has been and will remain critical to the advancement of AI.
Together, OSS developers will use AI to help make our lives better. I have no doubt that OSS developers will help build AI innovations that empower those with disabilities, help us solve climate change, and save lives.
As a European, I love seeing how open source AI innovations are beginning to break the narrative that only the US and China can lead on tech innovation.
I’ll be honest, as a European living in the United States, this is a pervasive—and often true—narrative. But this can change. And it’s already beginning to, thanks to open source developers.
Look at stable diffusion as an example. You’ve all likely seen exhibits of stable diffusion on the internet. Stable diffusion is an incredibly complex, deep learning text-to-image model that you can use to generate beautiful, lifelike corgis.
This corgi is not real! But you could fool me. Jokes aside, stable diffusion and models like DALL-E are one of the most profound AI breakthroughs to date. In August of last year, it was released by researchers at Ludwig Maximilian University in Munich, supported by an open source collaborative and German non-profit, two companies—and open sourced on GitHub. It’s a brilliant example of how OSS can help democratize AI. Originally, it was trained on thousands of GPUs at a cost of over 500,000 euros.
Then open source developers got to work, and their collaboration delivered huge efficiencies.
OSS built upon it, and made it better. Originally, you could only operate it on a high-powered device, and now it can be run on your smartphone—and yes, it can be spun up in seconds right on GitHub.
Now you can generate a corgi, or maybe a corgi with a rainbow? Wherever you go. And all that started in Europe.
Beyond flashy things like stable diffusion, the OSS community is building AI innovations that are testing for biases, keeping us secure, improving transparency, and so much more. The convergence of open source and AI is delivering incredible innovations, with impact.
Beyond this, open source injects enormous value into the economy with contributions that are especially important in moments like today. According to the OpenForum Europe study in 2021, that value is being delivered by some of the people in this room. Open source software is contributing up to 95 billion to the EU’s GDP. That is as much value injected into the EU economy as both the air and water transport industries combined. And that was commissioned before this wave of open source AI innovations began to take off.
According to estimates, AI is set to contribute over 15 trillion to the global economy by 2030. So it is almost unimaginable the economic value European OSS developers could bring by leading this wave of AI washing over the world. GitHub is the host for much of this AI innovation, and we look forward to measuring the economic impact of generative AI in the coming year.
Which is why, as I said earlier, the AI Act is so crucial. This policy could well set the precedent for how the world regulates AI. It is foundationally important. Important for European technological leadership, and the future of the European economy itself. The AI Act must be fair and balanced for the open source community.
Policymakers should help us get there. The AI Act can foster democratized innovation and solidify Europe’s leadership in open, values-based artificial intelligence. That is why I believe that open source developers should be exempt from the AI Act.
Because ultimately, this comes down to people. The open source community is not a community of entities. It’s a community of people. The compliance burden should fall on entities shipping products. OSS developers are often volunteers. Many are working two jobs. They are scientists, doctors, academics, professors and university students alike. They don’t usually stand to profit from their contributions. And they certainly don’t have big budgets and compliance departments!
The open source community act as the grassroots of technology, every day translating their ideas into shared possibility for the rest of the world. And the greatest innovations will always come from the roots—but only if we give them the daylight to grow.
Open source is forming the foundation of AI innovation in Europe. The US and China don’t have to win it all. Let’s break that narrative apart! Let’s give the open source community the daylight and the clarity to grow their ideas and build them for the rest of the world! And by doing so, let’s give Europe the chance to be a leader in this new age of AI.
Thank you so much for having me here today. Thank you to those watching online.
And to those here in the audience, I look forward to seeing you all become familiar faces in the years ahead.
Written by
Fascinated by software development since his childhood in Germany, Thomas Dohmke has built a career building tools developers love and accelerating innovations that are changing software development. Currently, Thomas is Chief Executive Officer of GitHub, where he has overseen the launch of the world's first at-scale AI developer tool, GitHub Copilot -- and now, GitHub Copilot X. Before his time at GitHub, Thomas previously co-founded HockeyApp and led the company as CEO through its acquisition by Microsoft in 2014, and holds a PhD in mechanical engineering from University of Glasgow, UK.