Video produced by Steve Nathans-Kelly
At Data Summit 2018, Hadley Reynolds, co-founder of the Cognitive Computing Consortium, presented a keynote looking at the meaning of AI and cognitive computing.
In particular, he considered what each of the terms AI and cognitive computing really mean and how they differ.
By 2016, IBM was starting to use the term AI, and in 2017, IBM "really" used the term, and so now everybody is talking about AI, said Reynolds. Given this, he noted, some might wonder why the consortium is still talking about cognitive computing.
“This is what we propose from the standpoint of our conversation,” said Reynolds: “That the fundamental differentiation is the extent to which the machine can emulate human thought processes, behaviors, and interactions. So, on the left under the AI umbrella, we propose that only fully autonomous systems should be considered in the realm of AI. Now, almost no system that you've heard recently tagged as AI is a fully autonomous system.”
Under the “autonomous” model, the machine is literally doing the work of the human brain, said Reynolds. “No human brain need apply. That's the beginning of autonomous. And then machine is literally put in the position of being author of its own actions.
Within the “cognitive” category,” said Reynolds, the machine behavior is dependent on the gray matter of the human that sets up the machine. “And the machine is an information tool and it's an agent of some vicious process or some intention of a human being and it becomes more like an assistant than an independent actor. In fact, it is, well, it's not necessarily an assistant, but very much can be.”
As a result of the confusion about terminology surrounding AI and cognitive computing as well as other terms such as machine learning and deep learning, and others, Reynolds says he has dubbed this period of time “the early chaotic era.”
There is confusion from multiple vendor interpretations, we're missing trusted resources, and credible guidelines and information are in short supply. “That's one of the main reasons that we founded the Cognitive Computing Consortium,” said Reynolds.
“The skill sets to deal with these technologies are in short supply. We're missing the cross-silo networks of people actually within organizations because the technology group and manufacturing group, and technology group and the marketing group, or maybe all three, should be talking to each other in ways about these cognitive ... I'm going to keep calling them cognitive applications. And they need to be able to talk together in order to generate the kind of decisions that would be made if a human committee from those groups were to make such a decision.”
To access more Data Summit 2018 videos, go to www.dbta.com/DataSummit/2018/videos.aspx.
Many PowerPoint presentations from Data Summit 2018 have also been made available for review at www.dbta.com/DataSummit/2018/Presentations.aspx.
Data Summit 2019, presented by DBTA and Big Data Quarterly, is scheduled for May 21-22, 2019, at the Hyatt Regency Boston, with pre-conference workshops on May 20.