How we can avoid an AI ice age

Parallels can be drawn between the mad rush to embrace artificial intelligence (AI) today and the hype around the Internet of Things (IoT) two decades ago. Both have enjoyed bold revenue projections, promised to change our lives and caused a stir with policymakers attempting to protect us from the risks of these “new” technologies. But can AI deliver on its vision to transform the world, or will it falter like the IoT? 

The IoT from its inception claimed it would revolutionize how we live and work. Referred to as the “fourth industrial revolution,” it was going to ubiquitously connect every “thing” into a single, intelligent ecosystem. There were forecast to be over 50 billion IoT devices by 2020, yet in 2024 there are only an estimated 17 billion. Companies such as Google and IBM are retiring their IoT solutions, and Cisco has estimated that three quarters of IoT projects have failed. 

The IoT has fallen far short of its potential for one reason — a lack of trust.  Trustworthy technology is built on two pillars — privacy and security. McKinsey’s recent research underscores the pivotal role of trust in IoT purchase decisions, attributing delays in IoT adoption to a trust deficit stemming from the inadequate convergence of the IoT with privacy and security. AI will suffer the same fate if trust is not built in at its foundation.

Though seemingly new, AI was birthed in the 1950s but has not been embraced until now due to a lack of compute power to enable it. AI’s early years delivered some successes including a breakthrough in artificial neural networks; however, it has faced two “AI winters,” one in the 1970s and another in the 1980s. I recall developing AI languages and linguistics while still at university and see today that these fields are considered by many as new. Now in its next favorable cycle and with compute limitations addressed, there is a real opportunity to turn the promise of AI into a reality.

Last week Apple revealed its approach to AI. The part of its strategy that stands out the most is that it has prioritized protecting personal data, addressing the key concern consumers and governments have with new technologies. As well, it has taken a measured approach, announcing AI use cases that can be adopted and add value today, rather than delivering only a vision of a grand imagined future. 

There have been some valuable applications of the IoT over the years, but the vast majority of the promise of the IoT remains untapped. The IoT’s most beneficial use cases are the ones that we have failed to bring to life, from transforming the health care industry to revolutionizing urban living, and many more advanced applications that have not yet been explored or defined.

If the technology industry puts trust at the core of AI, there is no doubt that AI will change the world for the better. Leaders across the entire AI value chain have a responsibility to build trust into their technology solutions for humankind to benefit from its tremendous potential. Otherwise, we will remain on the fringes of AI’s promise, as we have with the IoT, and perhaps, this time, face an AI ice age.

John Chen served as executive chair & CEO of BlackBerry, and chair, president & CEO of Sybase.