Why You Struggle to Explain AI’s Value
In 1994 while a student at Stanford I stumbled upon Mosaic Netscape — the first mainstream web browser. I loved it. I was convinced Netscape would change the world and couldn’t stop talking about it. I built a Pulp Fiction fan page— a handful of <Hx> tags — as my first “web site.”
I soon discovered how few people shared my enthusiasm. Sure, there was the Silicon Valley hype but most business people didn’t get it.
Me: Have you tried Netscape?? This Internet thing is going to be HUGE!!!
Them: I’ve heard of it. What is it?
Me: Well it’s going to change EVERYTHING. Imagine all of the world’s information connected and accessible at your fingertips.
Them: Ok. What can I do with it?
Me: Well…I made this cool Pulp Fiction fan page ALL BY MYSELF. Now anyone can get to it from any computer with Netscape.
Them: <<<<blank stare>>>>>
Talking about AI in 2017 is like talking about the Internet in 1994 or personal computers in 1975
95% of business people can’t relate to AI because the capabilities are too general while the examples are too specific.
AI’s relevancy gap
AI is an abstract general capability — like personal computers, the Internet, or social media — which doesn’t mean anything to most people.
Worse still, many actual use cases are so specific they trivialize the bigger picture.
Between these two extremes is AI’s relevancy gap: until people can see how AI will change their own business they won’t get excited about it.
Do you want to categorize images or sort cucumbers?
Computer vision is one of the most promising early uses for AI. Images are information-rich data sets which can be inexpensively (relative to sound, for instance) easily labeled.
Unfortunately the potential applications are also difficult to imagine because of the AI relevancy gap.
Too abstract: With AI you can categorize or score images.
Too specific: With AI you can sort cucumbers.
The cucumber-sorting application isn’t a joke. A Japanese farmer created the solution because his mother spends 8 hours per day sorting cucumbers by hand.
Automatically sorting cucumbers in 2017 is as interesting as my Pulp Fiction fan page in 1995: not very, unless you can see the potential impact of widely — available, powerful technology.
Jeopardy-champion AI gets the headlines. The real money is in sorting cucumbers.
We’re still in the early days of AI and hype (plus VC dollars) is getting ahead of the applications. But companies are making real money and our path forward is clear.
It will just take a year or two to plow through AI’s relevancy gap.
