Research Interests

Machine Learning and “AI” are transformative technologies and we have already seen major breakthroughs powered by deep learning systems. However, as with all transformative technologies it poses a number of risks too. My research focuses on minimising these risks. I have a background in mechanical engineering an optimisation and hence I am drawn to technical problems in this space. Research questions I am especially interested in include:

1. How vulnerable are frontier AI systems to attack and misuse by sophisticated adversaries?
2. How best to defend frontier AI systems against attack and misuse by sophisticated adversaries?
3. How can we reduce the cost of frontier AI systems on the planet?
4. How can frontier AI systems be used to tackle the problems of today, such as mis- and dis-information, climate change and inequality.

While the above questions are all “how” questions, I believe “should” questions are just as important and neglected by too many AI researchers.