In the 1980s, Professor Nicholas Jennings and Professor Michael Wooldridge began working on the problem of how autonomous AI systems might work together. Their work was crucial in establishing a field that ultimately led to Siri, Alexa, and Cortana.
Nick Jennings is the Vice Provost for Research and Enterprise and Professor of Artificial Intelligence at Imperial College London. Michael Wooldridge is Head of the Computer Science Department at the University of Oxford, and a Programme Director for AI at the Alan Turing Institute, London.
Congratulations on the award. What does it mean to you?
MW: The award is recognition that the idea of multi-agent systems - the whole bag of concepts, theories, algorithms and so on - is now a standard part of contemporary computer science. The award is recognition that we’ve reached that stage. For me, that is extremely gratifying.
NJ: I agree completely. For me, it's great that I've won it with my good friend Mike. I'm delighted also for the teams of researchers I’ve worked with: I've worked with over a hundred PhD students and postdocs along the years. Its wonderful recognition for all the help and ideas that they've contributed.
Back in the 80s and 90s, was there a grand vision when you initially started thinking about multi agent systems?
NJ: We saw that the world is not composed of individual actors that just act in silos: they interact to achieve individual and collective goals. I think the next big wave of AI, that will become fashionable, is people wanting to connect AI systems together. They will want these systems to cooperate, coordinate and negotiate with one another. Those are multi-agent systems.
MW: But it took a while! Around 2000, I was beginning to wonder whether we were barking up the wrong tree. But the dream was realised. We now all carry around a manifestation of that dream with us: Siri, Cortana and Alexa. These are agents. They're exactly what we were talking about. And the idea of Siri is that it's not just a dumb piece of software that you tell exactly what to do, it's an assistant. It's a personal digital assistant - to use the phrase that was used in the early 90s.
The reason it took so long for that dream to be realised was we didn't used to have the computing power in our pockets. Now we've now got supercomputers in our pockets.
But, the bit of the picture which we're just coming into is the idea that my Siri talks to your Siri. That part of the vision, which was always central to what we thought was going to be the future, we haven't yet seen pan out. But, I'm still confident that, that will be the future of AI. That's how the person on the street will interact with these ideas that we've been dreaming about over the last three decades.
How do you think your work will influence tomorrow’s world?
MW: Connected cars are another little piece of the dream of multi-agent systems. When we get to driverless cars, those cars will likely communicate with each other. Imagine a busy road intersection: what kind of protocols can you have for these cars that will allow them to maximize throughput on those busy road intersections? The kind of models that people have used include auctions to auction off slots - literally time slots on a busy road intersection.
BCS and its members are very active in the AI and ethics debate. Talk to us about the societal implications of your work.
MW: Accountability lies with people. You can't delegate your responsibilities and your legal obligations to a piece of software. It's humans that have to be accountable, humans that have to be ethical.