News
Why AI Needs Curious Minds
Engineering Manager at Google DeepMind, Brighton Girls’ alumna, Jennie Lees joined Oxford High School’s TEDx event to share her thoughts on AI. Below, we share an extract of her speech, edited for continuity.
“The game of Go is a strategy game which requires creativity, intuition and foresight. It’s similar to chess, but because there are a lot more options for any given move, it’s much harder to write a computer programme that can play it. This meant that for a long time, it was considered a ‘grand challenge’ in artificial intelligence. Twenty years ago, I was an undergraduate studying computer science. I got curious about this challenge and thought, ‘how hard is it?’. Well, the AI I created didn’t win a single game against a real person, and I concluded that the answer was ‘impossible’. Thirteen years later, an AI system called AlphaGo beat Lee Sedol, widely considered the greatest Go player of that decade. It was a pivotal moment for AI – the impossible had been achieved.
Today’s world presents tough, complex, global problems like climate change, disease, poverty, and inequality. These can seem impossible, but I believe that there is hope; that artificial intelligence can – and will – directly help with all of these challenges.
It’s already helping advance science. Every single protein has a unique 3D shape, the knowledge of which allows us to study the protein in depth, to better understand what it does and how it interacts with other molecules. The problem is, it’s expensive to figure this out – it can take a scientist years just to study one protein – and there are over 200 million to discover.
This is a perfect problem for AI to tackle. Scientists have worked out enough protein structures to kickstart a machine learning process, which is particularly good at going from a set of known things to a larger set of unknowns. Real-world science is more complex than Go, though, so nobody knew if it would be good enough to predict protein structures. In 2018, DeepMind unveiled AlphaFold, which took the 100,000 proteins whose structures were already known, and accurately worked out how to predict others. By 2022, it had published the shape of all 200 million, including every protein in the human body – another impossible problem solved.
It is hard to overstate what this single development means for science. Researchers are already using these AI-generated protein structures to investigate treatments for early onset Parkinson’s Disease. They are creating enzymes that can break down plastics to reduce waste and help the environment; and they are developing vaccines against diseases like Malaria. So, AI developments like AlphaFold can potentially save and transform millions of lives.
Other AI systems are helping to reduce the climate impact of air travel, save endangered species, and impact poverty by improving agriculture. None of these developments happen without humans who are inspired to challenge the status quo, to do impossible things and question what else is possible.
There are women and nonbinary visionaries, challengers, artists and leaders doing just that, bringing their diverse perspectives to the world as we know it today and redefining the boundaries of possibility. For example, Dr Kerry McInerney, a Research Fellow at the University of Cambridge, focuses on the intersection between artificial intelligence and society. She has uncovered ways that cultural bias can be echoed through AI systems, and has helped shine a light on gender inequality when it comes to AI in films. It turns out a higher proportion of onscreen scientists are portrayed as men than exist in our workforce today. For once, fiction is stranger than fact!
We also see role models in leadership who are bringing the power of AI technology to the world. Mira Murati, Chief Technology Officer at OpenAI (the company behind ChatGPT) has been at the forefront of the recent explosion in large language models, bringing technology out of the lab and into the spotlight, and helping ChatGPT become the fastest growing app in internet history.
If someone tells you something’s impossible, ask, “where do I begin?”
I don’t know what tomorrow will bring for AI, and I’m excited by that. A single AI system like AlphaFold will change science for years to come, and there are so many more problems we haven’t solved yet. Anyone can take an idea and have an impact on the world, more so now than ever. That might seem unreachable from where you are today, but I think the most important thing is to stay curious – and if someone tells you something’s impossible, ask where do I begin?’”
Jennie Lees
Jennie Lees is a software engineering lead at Google DeepMind, one of the world’s leading artificial intelligence research labs. An alumna of Brighton Girls, (Class of 2000), Jennie went on to a Computer Science degree at the University of Cambridge, and then her MPhil in Computer Speech Processing.
Moving to California, the epicentre of AI, she spent over a decade both I big tech companies and start-ups, as Engineering Manager at Riot Games, and as Director of Engineering at BossAlien.
She is a regular speaker, passionate about bringing new understanding to the positive potential of AI, and an ardent advocate for minority groups – in particular, women and girls – in the gaming and tech industries.
GDST Life Alumnae Magazine 2024/25
Jennie is featured in our 2024/25 edition of GDST Life alumnae magazine, where you will find a whole host of features and articles including stories, tips and viewpoints from a range of alumnae contributors, GDST and school news, our latest alumnae book listings, and how you can keep in touch.
