Why don't you just buy a refrigerator?
How the wisdom of the Three Stooges can inform the coming impact of AI
I’ve been thinking about what to write as my first real post here on Substack and I decided to start with the most important question facing lots of industries right now but two in particular that I really care about (and have some experience in) are learning and development (L&D) and innovation. Of course, because these are important topics, I sought wisdom from the sages known as Moe, Larry, and Curly.
Now the still above is from a 1941 short called An Ache in Every Stake. The Stooges have to deliver ice to a house at the top of a long flight of outside stairs. While Moe goes up to talk to the lady/client, Curly gets a big block of ice out of the truck and runs up the stairs with it. By the time he gets to the top, the big block is a little cube. Moe sends him back down and Curly, applying Stooge’s logic, gets two blocks of ice out and runs back up. Guess what? He’s left with 2 little cubes. At this point, Moe turns to the client and says “lady, why don’t you just buy a refrigerator?” Moe, without knowing it but with the Stooges’ genius, describes what I think is one of the central questions that #GenAI should be forcing L&D and Innovation Management to confront; what is our value vs our activity?
On the innovation front, so many of the models are concerned with that top of the funnel - the initial idea generation and idea assessment stages. That’s fine but that’s an activity that I can see building an AI-powered agent to do, at scale and at speed. Let’s say there are still humans involved (there should be) but its gut checking and those efforts are able to deal with so many more ideas because AI has helped winnow the pool but also has helped people generate so many more ideas. So what’s the value add? I’d argue that it’s in a couple of places - #1 out beyond the top of the funnel - horizon scanning, looking for signals and drivers, making orthogonal connections in a way that’s tough for AI to pull off. #2 deeper inside the funnel - down where the mass of ideas meets existing roadmaps and departmental budgets. Down where change management and organizational design have to kick in and are hopefully augmented by a supportive corporate culture. That’s hard work. That’s person to person work. There’s also huge value wrapped up there - ideas are cheap, execution is expensive. Read The Idea Factory: Bell Labs and the Great Age of American Innovation. One lesson there is that while its cool to have Claude Shannon riding his unicycle down the hall inventing information theory, it was really cool to have managers who took his ideas and crafted them into products and services we still use today. That’s value.
On the L&D front, the question of value vs activity, actually carries with in an existential concern in some instances. I say existential gently and with the utmost respect for what a strong L&D capability can bring to an org BUT I also know, from spending more than a few years in the space, that so much of what L&D teams are called on to provide, can be classified as an activity with limited value. I’m looking first at you compliance training - and understand me, compliance training is critical but the way we have been called on to provide is superficial at best - it is checking a box and doesn’t build any kind of positive pattern. So what happens when someone can develop an AI-enabled agent that will develop all the ethics, sexual harassment, and infosec awareness training content, customized for your org and featuring the latest content with voiceovers, animations, and assessments and do it all for a fraction of the cost in a fraction of the time. And I don’t mean like 1/2 the time/cost - think more like 1/50th. What then? When an LMS integration means introducing one AI to another AI and then watching them build all the APIs and required connections? What about when the AI (like ChatGPT) is working inside an environment that includes maybe the LinkedIn social graph and skills taxonomy, a huge content library, and is given compelling front ends like Viva and Teams and can generate personalized learning paths on the fly? What happens when your VILT class is taught by an AI-driven avatar (and the NPS/CSAT/Likert scores go up)?
So for both of these areas that I care so much about, that I think are so critical both to orgs being successful and the people in them being happy…I’m deeply concerned about what answer we will give when someone turns to us and asks….why don’t you just buy a refrigerator?
Much of our L&D lives are still hung up on that recurring notion that "we need to do better". But do we? Let's get this part out of the way: Yes, of course, we should always be striving to do better. And the digital assets or learning solutions we create should be powerful enough to chance behaviors and drive real change. Okay, now that we've dispensed with that, let's get back to reality. Compliance training exists for legal reasons. Legal issues are highly defined and detailed... lots of fine print. If a company can PROVE X, Y, Z, things, they will not be held liable for issues that would incur large financial costs. Solution: Hire person/team to do what legal doc says; no more, no less. And require that there is proof. That's all any company is concerned about. And that's why training courses are as mundane as they are. This is also why no company will spend more money to "make them better". Why would they. There is no such thing as being MORE compliant than you were before. You are compliant, or you aren't. Auditors will come in, review the documentation that proves everything is compliant. But they won't take your training course. Even if they did, they would say, "damn... that sucked...but you covered everything. You pass. See ya next year". Until the law says "prove that your training directly caused behavior change with a double blind study" you will never get your business leaders to buy a refrigerator. Small ice cubes get the job done.