Issue No. 05 - THE CLARITY MEMO: Unfiltered.

Non-Existent Studies and the Standards I Refuse to Compromise.

Founder’s Note: Leaders, this issue is a little different. Fresh off ChiefX LA Summit, where some of the sharpest executives were wrestling with exactly how AI is reshaping how decisions get made and where influence actually lives – I am compelled to reiterate that human judgment and critical thinking are not soft skills right now. They are your last defensible advantage. This is less a newsletter, and more a call to action.

While I was writing Issue No. 04, I came across a study describing gender-based disparities in how AI evaluates resumes. It was the kind of evidence that would sharpen my argument about the compounding effect of AI bias, and on the surface, it appeared unimpeachable. A major business publication had run an entire feature on it, and it had been widely cited and circulated across credible channels and institutions.

The only problem was that the study itself could not be found.

There was no sample size, methodology, participant demographics, geographic or cultural context. When I traced the research back to the person who claimed to have conducted it, the sole source noted on their website was the same publication article that originally covered it.

I did not use it. I pivoted instead, and that decision is the argument.

EVERY piece of thought leadership that leaves my desk; every newsletter, advisory brief, and strategic recommendation, goes through rigorous verification before it reaches a client or reader. It is a strict standard I hold because the leaders who trust my work are making consequential decisions with it in workforce strategies, AI governance frameworks, and boardroom arguments.

If the data underneath that work is remotely flawed, even if it comes from a credible-sounding source, or despite the fact that it’s been widely circulated; I have not helped anyone. I would have only handed them a liability dressed up as insight, and in doing so, I would have become exactly what I advise my clients against. In an environment where AI is generating content, summarizing research, and populating knowledge bases faster than any audit cycle can keep pace with, trust transfer is no longer an edge case and outsourcing critical thinking to the credibility of whoever shared it is now becoming the norm.

In the last issue, I shared research (credible and verified!) showing that when humans are placed in AI oversight roles, they concede to the machine's recommendations at alarming rates. I am extending that argument now, because the problem runs deeper than “humans are deferring to AI”. 

The humans who are actively reviewing, questioning, and fact-checking AI are often doing so without consistently deploying cognitive skills such as critical thinking, analytical reasoning, deep comprehension, creative problem-solving; that make that scrutiny meaningful.

These capabilities make human oversight real rather than performative, yet they are the capabilities being quietly deprioritized as industries race to close technical AI skills gaps. "A human was in the loop" is not going to hold up in a boardroom, a courtroom, or a public crisis if that human simply rubber-stamped what an algorithm – or an algorithm-assisted body of work – produced without examination.

Human presence is not the same as human judgment, and we must not conflate the two.

So here is what I am asking directly – hold the standard.

Slow down long enough to assess methodology and underlying strategy, over headlines and fancy decks. Ask where and how the original data was validated before building strategies on top of it.

Question the opinions everyone is citing, especially when everyone is citing it. The information environment is only going to get more complicated, and the volume of content with borrowed credibility is only going to grow. It is thus the fiduciary and ethical responsibility of leaders to ensure their judgment does not atrophy in the presence of AI tools.

Again, consider the fact that I almost cited a non-existent study, and the fact that it came wrapped in institutional credibility should concern you. Our standards, yours and mine, are the last real safeguard worth protecting; because borrowed credibility, absent critical thinking, and the human-in-the-loop illusion are compounding liabilities that AI governance frameworks are currently not designed to catch.

The integrity of any decision is only as strong as the judgment of the humans behind it.

The Clarity Memo: Unfiltered drops bi-weekly. Subscribeto get it directly to your inbox. And forward to the leaders in your orbit.

This content is for informational purposes only and does not constitute professional, legal, financial, or organizational advice. For guidance specific to your organization, contact Fadéké Strategic Consulting, LLC at admin@fadeke.com