Finding Gold in the AI Gold Rush: How Family Offices Can Separate Capability from Conviction
As family offices increase exposure to artificial intelligence, the hardest problem may not be finding opportunities. It may be determining which claims deserve to be believed.
Artificial intelligence has become one of the most convincing technologies I have ever worked with.
That sentence is not necessarily a compliment.
Modern AI systems can produce language that is polished, confident, technically sophisticated and completely wrong.
NIST uses the term “confabulation” for this problem: generative systems can confidently present false or erroneous content, sometimes accompanied by fabricated logic or citations that make the answer appear more credible than it is.
That creates an unusual problem for investors.
The machine no longer has to look confused when it does not know.
It can sound certain.
And increasingly, the person presenting the opportunity may be using the same technology to prepare the analysis, research the market, build the financial narrative and explain why the deal is compelling.
This does not make the analysis worthless.
But it changes the standard of verification.
On Tuesday and Wednesday, I will be at the Prestel & Partner Family Office Forum New York at the Harvard Club, where more than 100 family offices and UHNW decisionmakers will gather in a private setting to discuss investment, governance and the future of family capital. On Wednesday morning I will join Richard Awni, Jared Kash and Rondil Gosine for “The New Tech Superpowers: AI, Space Tech, and Deep Tech in Discussion.”
The timing could not be better.
UBS now identifies AI as the leading investment theme among family offices, with 65% of those surveyed already invested somewhere across the AI value chain.
There is real gold here.
There is also fool’s gold.
And the difference will not always announce itself.
I have spent more than 26 years at NASA’s Jet Propulsion Laboratory working on artificial intelligence, automated planning, scheduling, mission analysis and autonomous systems.
In that world, a system is not trusted because it gives an impressive answer.
It is trusted because the assumptions are understood, the boundaries are known, the result can be tested, and the consequences of failure have been taken seriously.
That discipline is useful far beyond spacecraft.
When I am asked to look at an AI opportunity, I start with a few simple questions.
What exactly is the system being asked to do?
What evidence demonstrates that it can do that reliably?
Was the evidence produced by the seller, by an independent party, or by the same AI system being evaluated?
What happens when the inputs fall outside the examples shown in the demonstration?
Where does the system fail?
Who has verified the claims?
And if the answer is wrong, how would we know?
Those questions do not require a PhD.
They require discipline.
A family office does not need to recreate an engineering laboratory before every investment decision. But it should develop a habit of asking for the layer beneath the presentation.
If someone says a model is 95% accurate, ask: 95% of what?
If someone says the system saves 40% of the work, ask: compared with what baseline?
If someone presents a market study, follow two or three of the cited sources yourself.
If an AI-generated report names a customer, a regulation, a scientific paper or a competitor, verify that it exists.
If the investment thesis depends on one technical assumption, find somebody who is not being paid by the seller to challenge that assumption.
These are small steps.
They can expose very large problems.
One of the paradoxes of this AI cycle is that the tools are becoming better at helping us investigate opportunities at exactly the same time they are becoming better at producing convincing mistakes.
So, the answer is not to reject AI-assisted research.
The answer is to use it with verification behavior built in.
AI can search faster.
It can compare faster.
It can summarize faster.
It can generate hypotheses faster.
But speed does not turn an unverified statement into a fact.
Eventually, a human being still has to decide what deserves trust.
That is the conversation I am looking forward to having at Prestel in New York.
Not whether AI will create extraordinary value. It will.
The harder question is how serious investors distinguish genuine capability from beautifully presented uncertainty.
In a gold rush, enthusiasm is easy.
Assay is what tells you whether you actually found gold.
_____________________________________________________________________
About the Artificial Intelligence Institute
The Artificial Intelligence Institute teaches people across ages, grades and professions how to remain in command of increasingly powerful intelligence systems.
Its proprietary behavioral architecture trains learners to move from cognitive offloading to cognitive amplification, questioning, directing, verifying and controlling the technology rather than surrendering thinking to it.
Think First. AI Helps Second.
The goal: turn advanced intelligence into a personal superpower without surrendering human judgment, authorship or control.
Learn more: artificial-intelligence-institute.com
Follow the Frontier
Follow the Artificial Intelligence Institute on LinkedIn, Instagram and Facebook for Frontier News Desk reporting, field observations, AI-readiness insights and emerging developments in human-AI interaction.
LinkedIn: https://www.linkedin.com/company/ethicalai-institutes/ | Facebook: https://www.facebook.com/profile.php?id=100094622704133 |
Family First Foundation
The Artificial Intelligence Institute grew from more than two decades of Family First Foundation work with children, families, schools and communities.
Follow Family First Foundation for the human, family and community side of this work and follow the Artificial Intelligence Institute for the frontier of AI readiness and human amplification.
Comments