There seems to be growing consensus among AI leaders that the field could end up transforming all of science and technology at a rate previously thought unimaginable.
After Anthropic CEO Dario Amodei had said that AI could bring about a 100 years’ worth of scientific progress in just 5-10 years, Google’s AI Chief Demis Hassabis has echoed his thoughts. “I really do feel like we’re on the brink of a new golden era of discovery,” he said at a conference. “I think what we need is a lot more interdisciplinary science. So using AI and asking the right questions with domain experts, it’s, it’s almost limitless what its applications could be,” he added.
“And of course, AI itself is a scientific discipline which is improving all the time,” Hassabis added. “So (it can) apply today’s technologies directly to other fields, and (can) also continue to improve AI itself. And that as well is a sort of exponential improvement. So there’s a lot of progress to be made in just the next few years,” he continued.
Hassabis cited the example of Material Science — the study of new materials — which could be one of the fields to be impacted by AI. “We’re touching on nearly every area of science. GNoME is one of my favourite projects which is on material design. And I think material design has some of the same characteristics that we look for in a problem that’s suitable for AI. It’s a massive combinatorial space. You need to try and build a model that understands the physics and the chemistry of the natural phenomena. And then if you have a model like that, you can then maybe use it to do a very efficient search through that search space, and then find an optimal solution. And in materials I think that would be groundbreaking — you could imagine designing new batteries, or maybe one day discovering a room temperature superconductor has always been one of my dreams,” he added.
Hassbis should know a thing or two about science — he was awarded the Nobel Prize this year in Chemistry for working on an AI model that helps determine the structure of proteins. He seemed to be suggesting that much of scientific progress today remains stuck in silos — the best minds in physics, for instance, don’t necessarily regularly speak to the best minds in biology to figure out how their respective research could be related. But if we manage to create a powerful enough AI that’s smart enough as Nobel Prize winners in all fields, it could combine the different knowledge bases in interesting and unique ways to create new discoveries. And given how fast AI is improving — NVIDIA CEO Jensen Huang says that AI is growing much faster than Moore’s law — a new age of discovery in science and technology could be just around the corner.