The young woman speaks directly into the camera. She has brown hair, wears a blue blouse and pink painted fingernails. You can hardly tell that she is not a real person. “I did not record what you are seeing and hearing here,” she says. “I have never been in front of a camera for this either. This is a deepfake of me.”
Alexandra Wudel generated the video, whose speaker looks like her, using artificial intelligence (AI) and posted it on her Instagram page. “This example shows what is already possible with technology today,” she says. “It is crucial that we use it responsibly.”
Wudel has also made this her own task. The 28-year-old business graduate heads the “Center for Feminist Artificial Intelligence”, or FemAI for short, which she co-founded last year. The Berlin start-up wants to make artificial intelligence more social and fairer and thus bring a stronger feminist perspective to the world of AI.
For example, deepfakes particularly affect women. AI-generated pornographic fakes have recently spread rapidly online. The faces of well-known women are edited into porn films that go viral on social media.And discrimination also occurs in other areas. If AI-supported programs make decisions based on distorted data, women and other underrepresented groups are overlooked and disadvantaged. For example, banks use AI to grant loans. If the program uses data from the past, men are often assessed more positively because they usually earned more in the past. Researchers at the University of Bath in the UK have also just discovered that women pay significantly more than men when it comes to AI-based car loans.
Solutions for large companies
The “FemAI” team checks programs to see whether they are inclusive and secure. It also develops its own technical solutions that test other systems to see whether they are ethically questionable. The start-up is already working with large tech companies such as Google and Microsoft. It also wants to be a link between research, companies and civil society. At regular community events that are open to everyone, people who have so far been little represented in digital policy network.
Word has gotten around. At the “data:unplugged Festival”, which is one of the most important events in the data and AI industry in Europe, Wudel was recently named “AI Person of the Year 2024”. Last year she received an award from a large banking group (Santander), which honors 35 remarkable leaders under the age of 35 every year. She has always been interested in the future, says Wudel. The question of how AI can be used to create a fairer world first occurred to her after she read Yuval Noah Harari’s book “21 Lessons for the 21st Century”. The chapter on the potential and risks of artificial intelligence and how this affects coexistence made a lasting impression on her.
At the Nordakademie in Hamburg, she specialized in digitization and innovation in her Master of Business Administration. After her thesis on the ethical dimensions of the AI law, she worked in a consulting company for ethical AI. She also organized the online election campaign of the Green politician Emilia Fester, who was elected to the Bundestag in 2021 as the youngest member of parliament at the time at the age of 24. In recent years she has worked as a political advisor on AI guidelines for the EU, the Foreign Office and the United Nations. With colleagues from academia, including from the TU Berlin, she has published on power relations and feminist AI design.
After 2 years of AI policy consultancy and research.
Alexandra and her team reflected on the results. The Tagesspiegel writes about this process: “Alexandra was initially enthusiastic about the basic idea of the AI Act, but the result is unfortunately disappointing.”
This is why, in the summer of 2024, Alexandra has redefined the strategy of FemAI from a research-based think tank into a socio-tech AI Start-Up during the summer of 2024. With this new approach she aims to drive both technological innovation and social impact.
