AI is getting more and more important. Emerging AI technologies allow us to radically shorten drug development cycles, accelerate the design of new materials and predict migration flows. But how do machines learn, and how can AI help us to speed up scientific discovery and innovation? Someone who knows is Alessandro Curioni, Director of the IBM Research Lab in Zurich, and an internationally recognized leader in the area of high-performance computing and computational science.
Curioni is a co-host of the conference Discovery on Steroids: How AI Will Speed up Innovation, organized by the Gottlieb Duttweiler Institute, the Swiss Re Institute and IBM Research on 5 July 2022. Prior to the event, he answered our questions.
GDI: Artificial Intelligence promises to revolutionise and speed up innovation. What does AI-powered innovation mean for society and businesses?
Alessandro Curioni: AI has been instrumental in driving progress and innovation for years – and with the ongoing research, it is getting ever more sophisticated. Consider, for example, the discovery of new materials. It’s always been an extremely tedious and expensive process, taking many years and considerable investment. We are starting to see how AI is infused along each step of the discovery process – from ingesting vast amounts of existing knowledge about materials to generating hypotheses about most promising new materials candidates to, finally, proposing and validating viable synthesis paths. And, perhaps most importantly, the integration and automation of all of these steps to achieve a virtuous cycle of scientific discovery.
When will we reach the point where machines make more decisions than humans?
It’s not about making more decisions – but about machines working alongside humans, complementing and augmenting our efforts. There are certain tasks that can be automated, thus much better suited for a machine – freeing up the human for other tasks and creativity. And there are tasks that a machine can do more efficiently. But at the end of the day, it’s all about making our lives easier with the help of AI – not creating AI that will outsmart people.
And when will we go from “Narrow AI”, that can perform specific tasks, to “Broad AI”, that is able to traverse multiple tasks and domains, or even to “General AI”, that is capable of complex reasoning and full autonomy?
We are working on that. There are efforts underway at IBM, in collaboration with other research institutes and academia, to create ‘common sense AI’ – meaning machines that reason more and more like humans. There is a huge effort at the MIT-IBM lab, for example, together with computer scientists, psychologists, cognitive scientists and other experts to create AI that makes decisions like human infants and the decisions become more and more complex with time – as the AI ‘learns.’ This is an incredibly exciting area of research, and we are definitely in the Broad AI era, where most of the business value is going to be created.
Regarding AI research and developments: What do you think will be possible in the future that’s unthinkable today?
In the future, synergies between AI and Quantum Computing could increase the capabilities and impact of both technologies in a very transformational way, something we can only start to grasp today. We may be able to simulate and understand our world in an unprecedented way, drastically accelerating the process of scientific discovery and the solution of key problems of our society.
Will AI be able to solve problems like climate change, food shortages, pandemics, and wars?
If used and implemented in a sustainable way, it will definitely keep getting better at helping us with solving global challenges. AI can help greatly accelerate the scientific method – and science has proven itself again and again as a crucial, indispensable tool humanity has to rely on to address the global challenges facing society all over the world.
A conference organized by the Gottlieb Duttweiler Institute (GDI), the Swiss Re Institute and IBM Research.