Dr Begüm Genç fills us in on her areas of AI research, the importance of ethical design, and how being a woman in this field is ‘not always easy’.
Dr Begüm Genç is a postdoctoral researcher working in the Insight Centre for Data Analytics at University College Cork’s School of Computer Science and Information Technology. She is also part-time lecturing in the same school.
Genç studied computer engineering at the Izmir University of Economics in Turkey, before completing her MSc with a focus on information visualisation for biological pathways at Bilkent University. In 2019, she obtained her PhD working in the optimisation and decision analytics research group at Insight.
More recently, she became the Ireland ambassador for Women in AI, a global organisation looking to close the gender gap in this area of tech.
‘There should be more focus on explaining to the end users the reasons behind decisions taken by AI systems’
– DR BEGÜM GENÇ
Can you tell us about the research you’re currently working on?
I am mostly active in two areas: algorithm design and AI ethics. The former emanates from my passion for designing AI models for quickly, correctly and responsibly solving problems that a human would need months or years to solve to the same quality.
The latter focuses on assessing different AI applications – such as AI in healthcare, intelligent cities, AI in human resource management – and identifying risks and working with the developers on how to address the underlying issues.
Some problems I look at can be found in many places in our daily lives, from delivering parcels to customers to creating ideal timetables at schools. Responsible development on these problems can take different forms such as explaining why a solution is presented as the best one, ensuring the proposed solution does not discriminate against any group, that it is robust to unexpected events, etc.
In your opinion, why is your research important?
I strongly believe that AI should be developed only to be used for the good of people. It should enhance people’s daily lives by carrying out the most tedious tasks and should not harm people in any way. My research is focusing on exactly these aspects.
With the increasing access to resources, we see many self-taught developers designing AI products. It is heartbreaking to see that many developers are not familiar with responsible design principles. They are not designing to harm, but they are designing products that may potentially cause harm and they are not aware of it.
To me, it is essential to incorporate the ethical dimension in the design phase of an algorithm.
What inspired you to become a researcher?
Having been raised by a family of engineers, mathematicians and academics, I had a natural inclination towards research and science. When I was a child, my family was often referring me to the big old encyclopaedias for questions that I had. I was always encouraged to search, read and find answers to my questions.
Back then, this was a game to me. So I became more and more curious about everything. Later on, it became a passion and eventually my profession. One of my favourite activities as a child was to build circuits and other structures with my older brother.
What are some of the biggest challenges you face as a researcher in your field?
I consider myself very fortunate to have very supportive supervisors who have also acted as great mentors throughout my research career.
However, being a woman working in the AI field was not (and is not) always natural and easy. I often needed to ask for support from male colleagues to reiterate my words so that my ideas could be heard. Similarly, often I was the only woman in the class or in the office, which gives the sense of ‘not belonging’.
Then, when you get more senior in research in academia, other challenges arise. For instance, challenges around building work-life balance in academia is often underestimated. You often find yourself attending meetings, writing reports or projects and lecturing during the day, and doing research in the evening.
If you wish to prioritise your family time in the evening, or if you have some caring responsibilities, you will not be able to publish as many papers as some others. If you don’t publish as many papers as others, you will not be promoted for the positions that are already very rare in academia. The hiring process is often all about the CV building and numbers.
Are there any common misconceptions about your area of research?
Misconceptions in AI are very common and I am not sure if all of them can be addressed! This could probably be caused by computer scientists not being great at dissemination. We build great stuff, but we are usually shy to explain what it is that we built.
For instance, people often think that AI acts like the human brain – but in truth, AI is only a software that the developers are implementing. The systems are far from comprehending their surroundings the way humans do. They are not able to learn on their own and we still struggle to implement an AI that makes jokes or understands jokes!
What are some of the areas of research you’d like to see tackled in the years ahead?
There are plenty of AI applications these days that are well known to discriminate against certain races, genders, religious groups, age groups, etc. In future, I would love to see AI evolving in the direction of eliminating the bias and discrimination factors to the best extent possible.
Additionally, considering that many day-to-day applications rely on AI models, which may result in wrong predictions, there should be more focus on explaining to the end users the reasons behind decisions taken by AI systems.
10 things you need to know direct to your inbox every weekday. Sign up for the Daily Brief, Silicon Republic’s digest of essential sci-tech news.