The Book I Have Returned To For Decades Says Your AI Problem Is Not Technical
Gerald Weinberg concluded that however technical a difficulty appears at the outset, it turns out to be a human one, and forty years have not improved on it.
Sign up for my Substack daily AI newsletter here.
See my AI Training course portfolio for corporate Business Leaders here.
Follow me on LinkedIn: https://www.linkedin.com/in/johanosteyn/
I read The Secrets of Consulting long before I had any business giving advice for a living. I have gone back to it many times since, and the return is never nostalgic. It is corrective. I run an AI advisory practice now, which I should state plainly, because what follows is a criticism of my own trade as much as of anyone’s procurement process.
CONTEXT AND BACKGROUND
Gerald Weinberg joined IBM in nineteen fifty-six and became manager of operating systems development for Project Mercury, the first American human spaceflight programme. He left in 1969 and in 1975 founded a consulting firm with his wife Dani, an anthropologist. He died in August 2013, having published, by his own count, around a hundred books.
He was, in other words, as technical as anyone in the industry, which makes his central conclusion harder to dismiss. Writing on his own site, he described what half a century with computers had taught him. People, he found, are at the bottom of just about every problem, and he added that he had known this as a child before being talked out of it, then spent years relearning it.
INSIGHT AND ANALYSIS
Every executive I meet has heard some version of this. Most agree with it in the abstract. Almost none accept it about their own organisation, and the reason is not stupidity. It is that the two diagnoses have very different consequences.
A technical problem can be purchased away. It has a vendor, a budget line, a timeline and a name. It can be delegated to someone whose job it becomes. A human problem has none of these. It cannot be procured, it cannot be delegated upward or downward, and it usually implicates the people commissioning the work. Faced with a choice between a difficulty you can buy your way out of and one you cannot, the pull toward the first is overwhelming, and it does not feel like avoidance. It feels like decisiveness.
Advisers make this worse rather than better. Selling a platform is easier than telling a chief executive that the obstacle is their leadership team, and it is considerably more lucrative. There is a version of the AI conversation in which everyone leaves the room satisfied, and it is the version where the problem was the technology.
IMPLICATIONS
For executives, one test is worth applying. When a proposal arrives describing an AI problem, ask what would have to change about how your organisation makes decisions for the proposal to succeed. If the answer is nothing, the proposal has not understood your situation.
For advisers, the harder discipline is to name the human problem early, before the relationship depends on not naming it. Weinberg’s own practice was to disturb an organisation just enough that its people became better at solving their own difficulties. That is a different business from supplying answers, and it is worse for revenue.
CLOSING TAKEAWAY
The technology I work with did not exist when Weinberg wrote. The organisations deploying it did, and so did their reasons for preferring a purchase to a reckoning.
That is why I keep returning to a book from nineteen eighty-five. Not because it predicted anything, but because it described the part of the problem that does not change, and told the truth about who is usually avoiding it.
Author Bio: Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net
Comments