Soumitra Dutta, AI and the fallacy of gradual change
"We all know that we are living in exponential times. What we have seen in the last five to ten years is nothing compared to what we will see in the next five to ten years. It's going to explode." — Soumitra Dutta, AI scholar and former dean of Oxford Said Business School.
It is likely not the march of technology, but rather society lagging behind the march of technology that will pose artificial intelligence's greatest challenge.
Companies and government bureaucracies have for decades organized around the assumption of gradualist change. Annual budgets, lengthy policy processes and education are failing to keep pace with AI, a rapidly developing technology.
As Soumitra Dutta, Oxford Dean cautions, the next decade will look nothing like the last; one can look at how quickly technological changes such as the internet and smartphones have become integrated into our society.
Yet for Soumitra Dutta, the warning is less about what is yet to come, than about the slow pace of change that organizations fail to overcome.
Soumitra Dutta, who for decades has charted nations' capacity to innovate through his Global Innovation Index and recently cofounded two AI startups, NexiVerify and Caasaa, has found time and again how essential adaptability is for businesses and governments.
His central message seems to be: innovation is as much about adoption and adaptation as it is about invention. And those are very different kinds of skill sets.
AI demands adaptation on a whole new level. Deploying the technology requires different skill sets, management practices, and governance models. Governments, moreover, must devise rules and adapt educational systems to prepare citizens and workforces for a world in real-time flux.
The focus is often on getting bigger, faster, more powerful chips, smarter algorithms and larger AI models, but that may be the easy problem.
The software will advance more quickly than human organizations can ever evolve.
Which is why Soumita Dutta's prognostication resonates. It isn't simply that AI will become smarter. It is that all human institutions organized on assumptions of gradualist technological change may find those assumptions outdated.
We will almost certainly produce more sophisticated artificial intelligence in the decade to come. Whether we will have better education, better government or smarter businesses remains a much more interesting question.
That might, all said and done, define how people benefit most from the revolution—not whether our chips are faster, but how quickly we are able to cope with this new reality.