Faster and safer AI adoption - AI is not "a thing"

April 10, 2024

As I talk to customers about AI technology I occasionally notice one barrier to its adoption: reification. But, if we take a different approach we can speed up adoption, have better safeguarding, and clarify things.

The sense that “AI” needs to be treated as a special concrete entity (i.e. reified), and that entity escapes the realm of normal product development or technology innovation, is a barrier. An unnecessary one. In this frame of mind, everyone is waiting for more definitions, clarifications, and unsure about how to treat this new “AI thing”. Risk averse organisations therefore will (understandably) wait. I covered how harmful some of this thinking can be - here.

Instead, if the frame of mind is of AI as a technology, the focus is to safeguard and benchmark the capabilities of the tools and systems that are produced with (or without!) this technology. Organisations can move faster by allowing the use of technology and prioritising a product or customer centric approach. For example, as a user of public health services, I want the safety checks on my treatment to be about its outcome, scientific rigour, safety etc. These safeguards should be in place no matter if the treatment is recommended by an AI system or by a human system. Also more on this - here.

AI is indeed a transformative technology that requires adaptability and change management. And of course there will need to be changes to regulations, and policies. So let’s not give AI more power than it needs - in the real world, and in our minds.