The new hype cycle

The Gartner hype cycle is a useful framing for thinking about new technologies:

Source: https://en.wikipedia.org/wiki/Gartner_hype_cycle#

 

Some key elements of the hype cycle:

  • Expectations inflate quickly and rapidly outstrip reality; the new technology is expected to transform the world

  • When people realize that expectations were inflated, the countervailing pressure creates a sense that the new technology is close to worthless

  • Eventually a balance is reached, somewhere in between the two. Often this represents the integration of the technology into existing human and organizational systems and processes.

  • The plateau level matters a lot. It doesn’t necessarily stay flat, but in fact usually rises over time, although often very slowly.

  • This pattern, loosely applied, is evident in ed tech examples including virtual reality, augmented reality, gaming, and other areas in the past decade or so.

AI is so big, however, that I wonder if it’s creating a different type of hype cycle—or will soon—in which the peak of expectations and the trough of disillusionment are going to operate simultaneously.

The hype is obviously happening all around us (along with real breakthroughs—it’s not just hype.)

But I think we’re starting to see, or soon will, the ways in which schools are limited in what they can actually do in the near term with AI because of existing policies, employment contracts, parent expectations, student privacy issues, and so forth.

And that’s going to lead to disappointment in the apparent capabilities of online tutoring systems, personalized learning, etc. I say “apparent” capabilities because tech enthusiasts may argue that these technologies will have far better capacities than they are able to show. I’m arguing that how they appear to many users will be disappointing because of the limiting factors—and perhaps because some systems won’t even adopt useful AI technologies in the near term.

Case in point, a recent conversation with a Colorado Division of Motor Vehicles chat bot:

 

This was a real conversation, and although there’s no date stamp I can verify that this conversation happened two weeks ago. Even in the moment I was more amused than annoyed at how poorly the chatbot performed. And I was briefly surprised that such a poor chatbot still exists.

I expect we’re going to see the same in public education as AI outstrips systems and processes, leading to a trough of disillusionment.

But at the same time, AI is likely to demonstrate incredible potential, keeping expectations high. Unlike the classic hype cycle, these two patterns may run in parallel, with the schools/systems/education sectors that are free of limitations able to demonstrate real and valuable uses of AI, while the policy-limited schools and systems mostly feel disappointed.

I do believe that the new plateau is going to be high, and highly disruptive. Probably not soon. A friend of mine who knows tech far better than I do told me a while back that the question in tech circles is “are we similar to Internet 1989 or internet 1999?” Meaning: are we two years away from an explosion of activity and change, or twelve years away?

An answer in education may be that we’re two years away for some, and twelve years away for others.

We welcome your comments in the DLC Community Portal's Blog Discussion Group. If you’re already a DLC user or member, you must log in before you can join or comment in the group. If you’re not yet on the DLC platform, please create a free user account or join as a DLC member to join the discussion.

Previous
Previous

100% Committed: Sustaining Inclusive Technology Practices in Online Learning

Next
Next

New uses of online learning