Earlier this week Eurasia Group published their Top Risks for 2018 which includes the risk of a global tech cold war. They write that, “achieving dominance in emerging technologies is the world’s most important battle for economic power.”

But what does “dominance” of an emerging technology mean? And what does that dominance look like? Does is mean only one country, or even one company, knows how to use that technology and fully unleash its potential? I can’t think of any examples of technologies that would come anywhere close to that definition today. And I have a hard time believing we’ll ever see a day that definition could be true. Certainly not in 2018. Sure companies have proprietary technologies and systems but alternatives are close substitutes.

The first cold war and the nuclear arms race that followed was driven in part by a race to supremacy. It was precisely that no country dominated nuclear weapons that propagated the cold war. And it ultimately took economic deterioration, not technological superiority, to cull the cold war.      

Eurasia Group argues “a race for breakthrough technology is underway between the US and China. Both countries’ tech giants are speeding to master AI and supercomputing among other highly investment-intensive, next-generation technologies. The winner could well dominate the coming decades, both economically and geopolitically.”  

Let’s break this down. First it’s misguided to suggest one can “master AI.” We need to stop talking about AI as if it is a single technology. There is no single killer application for AI out there that we are all rushing to discover. AI is not some Arthurian Sword in the Stone.    

AI systems today are incredibly narrow, precisely because companies are working on extremely diverse and differentiated services. While AI techniques like machine learning, speech recognition or computer vision can help improve efficiencies in numerous arenas, being able to apply these techniques effectively does not mean the gains are broad-based, cumulative, or exclusive.

The power of AI is in data. In most cases those data are proprietary to individual companies and their respective services or internal systems. AI fear mongers would have you believe AI is zero-sum, which it is not. Uber using AI techniques to improve time estimates does not preclude Lyft from using similar techniques to accomplish similar objectives. And It doesn’t obviate other companies from employing AI techniques in other service areas. We need not fear Uber or anyone else who finds useful applications of narrow AI techniques. It is a very far stretch to suggest success in narrow AI applications like improving time estimates could somehow lead to world domination.      

I’m also unconvinced that only a very few technology companies will benefit from AI and that access to AI will enable to them to lock-in their current supremacy. Corporate history shows companies have constantly lost their perch on the top while others have had to constantly reinvent themselves to remain relevant.    

In the spirit of Roy Amara, too many are overestimating AI’s current power and how quickly it will develop. And too many are overestimating the ability of gains in narrow AI applications to be widely applied. AI techniques will be broadly important to a wide range of businesses and we are just now discovering what some of those applications and use-case scenarios will look like.