Gartner, a consulting firm, developed the Hype Cycle to watch the diffusion and adoption of technologies[4]. They found that adoption of technology didn’t follow a nice bell curve like the one Everett Rogers predicted in his diffusion of innovation work (innovators, early adopters, early majority, late majority, and laggards). Instead, innovations grew in expectation, didn’t meet the inflated expectations, and then slammed out of favor, and eventually found a place where capability and expectation matched. They called it the Hype Cycle.
A "Technology Trigger” would gain some attention. That attention would grow to a “Peak of Inflated Expectation” then fall into a “Trough of Disillusionment,” followed by a growing “Slope of Enlightenment” and finally find a “Plateau of Productivity.”
When SLAAIT began, Generative AI was at the very tippy top of the "Peak of Inflated Expectations.”
Maximum hype. It is bound to fall when people with a straight face talk about how Generative AI will cure disease, solve the climate crisis, and even fix democracy. It won’t, at least not in its current state. It has already been proven that it will change how written work is done, evaluated, and treated in scholarly and literary circles. So, what will stick? Will there, in 5 years, be a Copilot[AI]+PC as just announced by Microsoft[5]…probably not. Any good technology that lasts begins to submerge behind the problems one is trying to solve, not take center stage looking for problems to fix.
So, is that it? Is SLAAIT done? We saw the rise, now we will document the fall, and look for the useful? We need to take a critical eye to the promises and look for the useful. However, we also need to step back and look at our terms.
On Gartner’s Hyper Cycle while Generative AI sits at maximum hype, it is surrounded on both sides by a host of different types of AI. Coming up the hill? Causal AI, Composite AI, AI Simulations. On the way down? EdgeAI and AI Maker and Teaching Kits. And all over the graph AI and machine learning related topics like Computer Vision, Autonomous Vehicles, Prompt Engineering, Decision Intelligence. Here we see inviable AI – systems at the core of the technology we do notice-is changing our world without the hype.
AI approaches have already transformed things like text detection that runs scanning projects. Where machine translation used to depend on complex mapping of languages, AI has done a faster and better job in translation…and in real time. Scanning texts, translating them, and speaking them aloud all have a real impact today on the work of libraries. And for our communities? AI has already disrupted how people are reviewed for jobs and hiring decisions[6]…and in granting parole[7], and in setting insurance rates[8]. In fact, while Generative AI has grabbed the spotlight, more and more “mundane” functions are being transformed by deep learning and neural networks.
We all need to keep this in mind as our work proceeds.
[1] https://www.linkedin.com/pulse/we-reaching-llm-plateau-david-elliman-frsa-fbcs-po8se/
[ 2] https://www.inc.com/jason-aten/mkbhds-humane-ai-pin-review-reveals-its-fatal-flaw-its-not-bad-tech-its-just-a-bad-idea.html#:~:text=A%20lot%20of%20reviewers%20have,Then%20there's%20Marques%20Brownlee's%20review.
[3] https://apnews.com/article/new-york-city-chatbot-misinformation-6ebc71db5b770b9969c906a7ee4fae21
[4] https://www.gartner.com/en/research/methodologies/gartner-hype-cycle
[5] https://blogs.microsoft.com/blog/2024/05/20/introducing-copilot-pcs/
[6] https://financialpost.com/fp-work/ai-job-interviews
[7] https://theconversation.com/a-black-box-ai-system-has-been-influencing-criminal-justice-decisions-for-over-two-decades-its-time-to-open-it-up-200594
[8] https://www.mckinsey.com/industries/financial-services/our-insights/insurance-2030-the-impact-of-ai-on-the-future-of-insurance
