Gaël Gendron

Gaël Gendron is a PhD student at the School of Computer Science whose research is focused on artificial intelligence, machine learning and causality theory.

At the invitation of Julie Rowland, the Deputy Dean of the Faculty of Science, Gaël recently participated in a panel discussion about the impact of AI on sustainability and higher education.

“The panel discussion was a tremendous opportunity to discuss topics I’m passionate about that have many societal implications with world-renowned experts. The room was packed, and I got positive feedback from the audience, so it was a great experience!”

Can you tell us more about the panel discussion?

“I was honoured to join a panel of impressive thought leaders, including Professor Siah Hwee Ang, director of New Zealand’s Southeast Asia Centre of Asia-Pacific Excellence, Professor Teck Seng Low, senior vice-president for sustainability and resilience at the National University of Singapore, and President Banchong Mahaisavariya of Mahidol University and moderated by Professor Julie Rowland. We debated and answered questions from the audience on the future impact of artificial intelligence on sustainability, how it could help or prevent us from achieving our climate change goals, and how it would impact our lives, particularly in information and education.”

What is your main takeaway from the event?

“I learned a lot about sustainability and how people worldwide are trying to tackle its challenges. I thought I would feel out of place as ‘just’ a student, but I realised that I had a lot to offer from my knowledge and perspective, and I wish to see more young people and students appropriate these questions and intervene in these kinds of events.” Gaël is completing a PhD in computer science (artificial intelligence, machine learning and causality theory) at the University of Auckland. He moved here from France to complete the Research Internship element of his Masters in AI degree and decided to stay on and complete a PhD.

“I was looking for academic excellence, and a high-functioning AI research centre in a desirable location – New Zealand and the University of Auckland ticked all the boxes.”

“I joined the Strong AI Lab, which had just been founded by Professor Michael Witbrock, for six months. I enjoyed the experience in every aspect and was in contact with AI experts who provided me with excellent supervision, so I decided to stay for at least three more years and do my PhD here."

What is your research/thesis topic?

“Improving the capacity of AI systems to reason and think more “human-like.” AI cannot currently form strong abstractions that can be reused in multiple contexts. They either reason in a narrow way: they are very good at a specific task but fail if you tweak a few settings; or, in a shallow way, they provide good answers on the surface but without proper understanding. For instance, ChatGPT can give you a confident answer on almost every topic, whether wrong or right, like a student who did not study and is bluffing in front of the teacher. I’m trying to solve this issue by making them understand cause-and-effect relationships and build proper causal reasoning structures when answering a question, which can reduce their tendency to confabulate.

What excites you most about artificial intelligence?

One of the most fascinating aspects of AI is the speed at which research progresses. If you look back five years ago, many of the AI technologies available today would have seemed like science fiction. If you extrapolate this success to five years in the future, you can imagine many incredible or worrying applications! One that fills me with concern and excitement is the creation of an ‘Artificial General Intelligence’ that would equal human thinking on most topics and could make new scientific discoveries that would otherwise take years for human researchers. However, whether or not this can be achieved is heavily debated. A more grounded application I’m particularly excited about is the use of AI for medicine. There are many ways AI can help, ranging from drug discovery (DeepMind’s AlphaFold has already considerably sped up research by solving protein-folding problems) to doctor assistance and personalised diagnostics (generative AI and causal models have already shown great promise).

Can you share any accomplishments or milestones you’ve achieved?

I’ve had four papers accepted to top AI conferences and one of them was awarded the “University of Auckland Best Student Published Paper in Computer Science.” My PhD topic has also been accepted for presentation at a doctoral consortium at a top AI conference.

Finally, tell us something about yourself that we can’t learn by Googling you.

I’m actually very shy! A few years ago, I would never have imagined that I’d be talking about AI in front of a crowd, let alone enjoying it!