To give AI-focused female academics and others their well-deserved – and overdue – time in the spotlight, TechCrunch has published a series of interviews focusing on notable women who have contributed to the AI revolution. As the AI boom continues, we’ll publish these articles throughout the year, highlighting important work that often goes unrecognized. You can find more profiles here.
Catherine Breslin is the founder and director of Kingfisher Labs, where she helps companies develop AI strategies. She has been an AI scientist for more than two decades, working for the University of Cambridge, Toshiba Research, and even Amazon Alexa. Previously, she was an advisor to the VC fund Deeptech Labs and a Solutions Architect Director at Cobalt Speech & Language.
She studied at Oxford University before receiving her master’s and doctorate degrees at Cambridge University.
In short, how did you get started with AI? What attracted you to this field?
I loved mathematics and physics at school and decided to study engineering at university. That’s where I learned about AI for the first time, even though it wasn’t called AI at the time. I was fascinated by the idea of using computers for speech and language processing, which comes easily to us humans. From then on, I ended up studying language technology and working as a researcher. We are at a point where AI has made great strides recently and I believe there is a great opportunity to develop technologies that improve people’s lives.
What work in AI are you most proud of?
In 2020, in the early days of the pandemic, I founded my own consulting firm with a mission to provide organizations with real-world AI expertise and leadership skills. I am proud of the work I have done with my clients on diverse and interesting projects, and also of being able to do this in a truly flexible way with my family.
How do you overcome the challenges of the male-dominated technology industry and therefore also the male-dominated AI industry?
It’s hard to measure accurately, but about 20% of the AI field is women. I also have the impression that the percentage decreases as I get older. For me, building a support network is one of the best ways to cope. Of course, support can come from people of any gender. However, sometimes it’s reassuring to talk to women who are in a similar situation or have seen the same problems, and it’s great to not feel alone.
The other thing for me is to think carefully about where to put my energy. I believe we will only see lasting change when more women get into leadership positions, and that won’t happen if women spend all their energy trying to fix the system instead of advancing their careers. There is a pragmatic balance between driving change and focusing on my own daily work.
What advice would you give to women wanting to enter the AI field?
AI is a huge and exciting field with a lot going on. There is also a great deal of excitement about the seemingly constant release of papers, products and models. It’s impossible to keep up with everything. Furthermore, no matter how eye-catching the press release, not every paper or research result will be significant in the long run. My advice is to find a niche that you really want to make progress in, learn everything you can about that niche, and tackle the problems you want to solve. This will give you the solid foundation you need.
What are some of the most pressing issues facing AI as it continues to evolve?
There have been rapid advances in the last 15 years and we have seen AI move from the lab into products without really taking a step back to properly assess the situation and foresee the consequences. An example that comes to mind is how many of our speech and language technologies work better in English than in other languages. This does not mean that researchers have ignored other languages. Significant effort has been put into non-English language technology. But the unintended consequence of better English language technology means that we develop and deploy technologies that do not serve everyone equally.
What issues should AI users be aware of?
I think people should realize that AI is not a panacea that will solve all problems in the next few years. It can be quick to create an impressive demo, but it takes a lot of effort to build an AI system that works well consistently. We should not lose sight of the fact that AI is designed and built by people, for people.
What is the best way to build AI responsibly?
Building AI responsibly means including diverse perspectives from the start, including from your customers and everyone affected by your product. It’s important to test your systems thoroughly to ensure you know how well they perform in different scenarios. Testing has a reputation for being boring work compared to the excitement of thinking up new algorithms. However, it is important to know whether your product really works. Then you need to be honest with yourself and your customers about both the capabilities and limitations of what you are building so that your system is not abused.
How can investors better advance responsible AI?
Startups are developing many new AI applications, and investors have a responsibility to think carefully about what they want to fund. I would like to see more investors express their vision for the future we are building and how responsible AI fits into it.