Roelof Pieters is one of our panellists in the afternoon. He is one of the leading AI visionaries and developers in Europe and the founder of Creative.ai and Sunshine Lab. He also founded the influential AI-network Stockholm AI. We invited him to tell us more about himself ahead of the conference.
Roelof, I’m fascinated by your background. You have a Masters in Social Anthropology and then became a PhD Candidate in Theoretical Computer Science, and now you’re the CEO and Founder of Creative.ai and Sunshine Lab. I think anthropology and computer science is a perfect combination but I’m intrigued how you came to follow this path?
I have always been intrigued by people *and* technology, mainly how amazing it is that we can extend our own capabilities to do more. I originally studied computer science but dropped out the last year because I found it too commercially-oriented (which is ironic as I’m now the founder of multiple companies myself). Instead I wanted to use my skills as a software developer to help the Global South, not as charity, but from a sense of shared community: we are all inhabitants of planet earth. For my final dissertation, just before I dropped out, I teamed up with a group of students from different universities and disciplines and with the help of a grassroots organisation, Midogo, we went to Ghana to set up an Internet café (this was pre-smart phones). But while the work we did, small and local in focus, came together pretty well, I wanted to further explore the ways in which my technical skills could benefit the world, and most of all how it might help fight some of the biggest injustices and inequalities in the world, as is still very much the case.
I wanted to further explore the ways in which my technical skills could benefit the world, and most of all how it might help fight some of the biggest injustices and inequalities in the world.
My then girlfriend told me about this discipline called cultural anthropology that seemed to match all my interests around development work, culture studies, studies of colonialism, post-colonialism, etc. I really fell in love with anthropology and ended up focusing on digital anthropology, as well as anthropology-at-home, mainly Europe and The Netherlands, and studied and wrote extensively about the rich history of Dutch social movements, mainly around environmental, housing, migration, and animal-rights struggles.
During this time I worked part-time as a software developer to pay for my studies, but it wasn’t until I moved to Stockholm for my Masters in social anthropology that I started to write software. This was also the time digital humanities started to be a “thing”, or at least when I started to be aware of it. The software suite could parse large collections of newspaper articles from lexis-nexus, basically thousands of articles around different protest groups, for my dissertation.
At the start-up where I was working during my studies, I led the R&D team for video analysis for a Netflix-like recommendation engine. That’s when I got very interested in AI, and especially the need to critically investigate such technologies. This was in 2013/2014 when Deep Learning just started to be “hot”. So I started a PhD in AI/CS at KTH in Stockholm. They didn’t really have a deep learning lab at that time so I started my own company, data science and consultancy, to be able to do an industrial PhD. It meant I had complete freedom to do my research, with the benefit of having office space and other facilities and research peers at the university. I did that for a couple of years until, through research in the inner workings of neural nets, I found that I could kinda put them inside out and make them generate things, rather than only do what they were up until then used for: classify images of cats and dogs. This was around the time that Deep Dream and some time after Style transfer would become the thing they are today: neural nets on LSD.
Can you tell us a bit more about Creative.ai and Sunshine Lab?
Yes, sure. Creative.ai was born out of a joint passion between Samim Winniger, Alex Champandard and I for doing creative things with AI. Samim and I ran a small consultancy doing weird arty projects like neural net LSD-like dreaming over a music video of the pop band Years & Years we were hired to do and such things.
We had some crazy ideas about democratising AI: making it more accessible to non AI researchers and having wider societal discussions about this technology. That and just having a crazy good time making neural nets do weird things. We wrote a blog post on what we imagined Creative AI, as we called it, could be, going back to the 60s/70s type of AI thinking that was very much forgotten, when AI was at a crossroads between the kind of singularity, Silicon Valley, white-male type of thinking that is mostly understood as AI these days, and a more progressive type of augmentation of the human intellect and capabilities by people like Douglas Engelbart, where AI is just a tool. But a tool that can be used to help solve large societal problems, and in doing so become a better tool able to solve even more difficult problems. The point being: AI as a tool that is ultimately guided by a person versus AI that starts to guide people. A question of agency basically.
AI is just a tool. But a tool that can be used to help solve large societal problems, and in doing so become a better tool able to solve even more difficult problems. The point being: AI as a tool that is ultimately guided by a person versus AI that starts to guide people.
So we decided to start creative.ai as a company to build tools for creative work and exploration.
Today we work somewhere between an agency or consultancy where we help creative industries with AI, and a tech start-up with some software-as-a-service type tools around video analysis for digital archives (think broadcast companies), and a tool blōma, which will be open source soon.
I have also founded a new company, Sunshine Lab, the thinking for which goes back all the way to my work in Africa, as well as creative.ai: using tech, and I really mean tech as a broad thing, not just AI. Increasingly I have come to believe that AI can’t really do that much in solving the really large problems we face today, like climate change.
AI is great for two things. One: Making some things perform better, it can optimize, etc. This is mostly what AI is used for today. Two: It can be used to analyse, make sense of things, the world, data. For most companies this means surveillance, to better target users with ads, products or services.
AI-for-good, as it’s commonly called, can also be used, for instance, to fight climate change, it can detect deforestation from satellite imagery, it can detect forest fires, illegal fishing, etc from satellite imagery. But someone will still have to go there and actually do the hard work of stopping forest fires from spreading, etc.
At Sunshine Lab we build products to help climate change. Climate change is the single most urgent problem of our time and we work on technical solutions that are not only human-centric as we love to say in design land, but planet-centred: technology that takes a more holistic look at things. We believe, for instance, that carbon reduction is but one thing among many urgent problems that are all interconnected. The United Nations Sustainable Development Goals are a great example of such a more holistic understanding for instance.
We build products to help climate change, which is the single most urgent problem of our time. We work on technical solutions that are not only human-centric…but planet-centred: technology that takes a more holistic look at things.
Currently we’re working on a home solar generator unit that is super lightweight, is the size of a small backpack, and can power an entire household in African countries. Not just some LED lights as is currently the standard but also, for example, a refrigerator that can run 24/7, even if the sun is not shining for 3 days. The challenge is more around hardware. But we’re also working on another project around bio materials where AI is actually very helpful. We take a cradle-to-cradle approach where all materials are maximally recyclable and we make sure raw materials don’t come from problematic areas.
Why are you excited to attend the conference and what do you hope to get from attending?
I’m super stoked to meet fellow tech-anthropologists, geeks, and technologists with a heart and a brain. I hope to meet many friendly and exciting people, to be inspired by some talks, and conversations, and perhaps build on the momentum that is currently building around climate action and such issues.
In your opinion, why do you think other technologists should attend the conference?
I — and I feel very strongly about this — believe that we all need to work together, learn from each other and especially embrace the social sciences as a key partner in all technical adventures, anthropology being one of the important ones.
I feel very strongly that we all need to work together, learn from each other and especially embrace the social sciences as a key partner in all technical adventures, anthropology being one of the important ones.
As well as being a panellist in the afternoon, you’ll also be giving a PechaKucha talk. Can you tell us a little bit about what you might be presenting on?
Ah, I don’t want to say too much at this stage! But it will probably be around some of the thinking and explorations I’m doing right now with Sunshine Lab about the role that technology and more specifically AI can and should play in society at this very moment, and the role technologists, but also designers, artists, and anthropologists — among others — might play in being a progressive force for change.
Is there anything else you’d like to tell us or say?
I am looking forward to seeing you and everyone in October!
Thanks so much, Roelof, we can’t wait to meet you on 3 October!