Top secret: the number one thing innovation leaders can do to give them an edge. Shhh!

When I advise C-suite clients and start-up founders on adopting AI solutions or any emerging technology operation, especially if I'm hired to build a solid innovation culture, I recommend learning in one subject area above all others. It isn't science and math.

It's history.

Doomed to repeat, and repeat, and repeat all over again.

From the C-suite down to the developer level, a solid strategic vision only emerges from knowing the past's rhetoric, experiments, and frauds. For example, let's take a subject in the news lately. Automation. "AI" driven automation. For a convenient example, let's take automation for automobiles promising driverless automation. Several cases have come up recently where human deaths and severe injuries have resulted from automated driving features, and some of them have settled. In all cases, there is something implicated that we've known about since the 1980s with concrete evidence and hypothesized about before that critical moment, most famously documented by Lisanne Bainbridge. It's called supervisory control.

It's so well-known and thoroughly documented that it is noted not only as a psychological phenomenon but also as a neuroanatomic one. We've documented pictures of supervisory control on MRIs, so we know its perils and implications in any operation that requires a person to maintain semi-vigilance.

This means that the person driving the autonomous vehicle or monitoring any automation can never entirely relax, leading to cognitive overload, which deteriorates performance over time. Further, the lack of engagement in the supervision is often dull and can lead to minds wandering. This was considered problematic in manufacturing. Not only does it sometimes lead to catastrophic errors, but it also creates mental stress for humans in supervisory mode. Some might argue that all innovation comes with risk and human death, and they wouldn't be wrong. However, as any sound engineer will tell you, when you plan, there are three categories of risk that you must address to succeed and be cost-effective:

1) knowns

2) unknowns

3) unknown unknowns

It is only 3 where the risk analysis provides benefit. In 2, there are ways to model unknowns on paper; depending on multiple variables, you can expect to uncover at least 85% of the problems on paper. Anyone who entails the risk of the issues in number 1, when the risk involves human life, is reckless at best and may be criminally guilty of manslaughter at worst.

Adverse human impacts involving supervisory control are firmly in category 1.

Beware the ghost in the machine.

Are you familiar with the long history of automation machines and fraud? The first recorded example was the Antikytheria Mechanism, widely accepted as the first documented computer dating to the first century BC and perhaps earlier. It was used to predict the movements of celestial bodies. Those who could demonstrate knowledge and power of the heavenly bodies gained a lot in ancient societies, not unlike our current fascination with space exploration on the one hand and not like data platforms that claim they can predict markets on the other. It is also widely believed to have been fraudulent and more of a work of art than a machine. Because, as Steve Jobs knew so well, the aesthetics of our machines have always mattered.

Similar examples exist throughout almost every period of history. Another common theme is human actors working undercover to fool the audiences of their machines into thinking they are working independently. The most famous was the Mechanical Turk, which inspired those carnival games with Voltar, like in the movie Big. The fact that they feature in fantasy and horror films so prominently shows that our fear and fascination with animated machines persists, to say nothing of Gen Y's fascination with FNAF. It also demonstrates that none of us are above magical thinking; in some ways, the human mind craves illusion and fiction. However, there is a difference between being deceived for pleasure, like with magic shows, and becoming convinced that fiction is real because we need to escape accountability and responsibility for being an adult.

The most famous example of the human behind the machine (like the Wizard of Oz) is from the 18th century and is called the Mechanical Turk. This combines not only our magical thinking regarding machines and needing to believe they can replace human action but also the embedded Orientalism that allows people to deceive people from the West with images of the East because of the Westerners' own prejudicial belief that they are somehow magical. Prejudice and systemic bias always lead to stupidity in every case.

When we see people investing in big and powerful fancy prediction machines, we must conclude either of two things: One, they don't know this history, so they are being profoundly misled. Or two, they know and are banking on this human weakness to deceive and consolidate wealth. There are other options, but I think they are subcategories of these two. The issue is with the second option; they may start with this notion, but there is another human psychological phenomenon where the deceivers begin to fall prey to their own deceit. So I'm skating over some complexity here to inform you that to have an edge like no other, embrace emerging technology with two eyes wide open.

Coda

And finally. Hire independent, solo consultants who do not depend on the large-scale sale of "solutions" for their income but will honestly advise you without a stake in adopting a platform solution. (It doesn't have to be me.)

What do I mean by solutions? If the answer to your problem is a platform services partner they recommend or a large platform solution that sits upon an API hub that promises things like "seamless integration," "velocity," and "automated data solutions," with pictures of rockets and astronauts and the wide-open sky, Gentle Readers, you are not working with a consultant; you are working with a salesman dressed up as an intelligent advisor. An illusion that falls apart at less than a millimeter deep.

If you want to learn quicker and fail faster, learn from the past. You will emerge miles ahead of your competition. If you want to get a valid consultation contract with independent, vendor-agnostic consultants who are transparent about the sources of their revenue.