Technology breaking paradigms
As if Artificial Intelligence wasn’t complicated enough, there’s now yet another term bringing more complexity. Applications of AI can be generally summed up into one of three categories – DIY, faux, and transformative. The first two are most common, and fairly self explianatory. DIY AI gathers information for a user that then carries out an action themselves, for example picking a restaurant out of a list of suggestions, and Faux AI is a touted as Artificial Intelligence but isn’t actually based on machine learning properties. Transformative AI, on the other hand, gathers, processes and acts on information without human intervention. In a scientific sense, transformative technologies break established paradigms.
This third category of AI application is the most disruptive class of Artificial Intelligence – the one that various commentators believe could take over the world. An example of transformative AI can be seen in autonomous cars. Instead of prompting a human to respond in a certain way, self driving systems can make the decisions. Many of the DIY algorithms used today are transformative AI in training, accumulating masses of data which will eventually contribute to independent decision making. The technology is already well under development. Companies like Transformative are harnessing the ability of self acting AI for healthcare, but it’s clear that any other industry could be disrupted too.
The growing vocabulary surrounding machine learning demonstrates the expansion of AI as a technology and a discipline. Understanding transformative AI is instrumental in insuring that it remains a tool rather than a threat. Non profit and regulatory bodies like OpenAI and the Partnership on AI claim to be focused on ensuring that all AI applications are benevolent and safe, so that if and when transformative AI becomes ubiquitous, it can be controlled.