SHRM26 Debrief · Leadership & Development · #2525

AI Made Work Easier. Did It Make Your Workforce Weaker?

with Ashley Miller

& Alexander Alonso

▶ Watch on the SHRM portal ~77 min, distilled

skill phishingAI governancehiring validation

"A person who prompts and accepts whatever the output is within six prompts is unhireable in my shop. Unhirable."

What it was about

AI is making it easier to produce polished-looking work and polished-looking candidates while masking real capability, so hiring, AI governance, and workforce development are no longer separate risks but one strained workforce capability system that HR must verify, govern, and rebuild.

By the numbers

78%

estimated actual share of workers using AI (personal AI) to accomplish work, versus the 41% who admit it

of CHROs say they are actually leading the human side of AI integration

3 months

estimated current time to identify an incompetent 'skill fisher' hire, vs. about 1 month twenty years ago

Key notes

Move hiring from evaluating declared skills to evaluating demonstrated skills using work samples, in-basket exercises, simulations, or at minimum structured behavioral interviews.
Probe candidates' AI use directly by asking to see their prompt history and asking what tools they use across the board, not just whether they used AI.
Preserve a deliberate human review step in every AI-assisted workflow rather than letting AI evaluate AI-generated output.

The contrarian takeWorkers over 45 are the most effective users of AI in the workforce today, not younger workers — because they bring critical thinking, agency, and experience that AI-native younger workers lack.

Take this back Monday

Do this for your team

Add one required question to your next interview: ask a candidate to show 7+ prompt iterations behind a work sample, not just the final output.

Say this in your next leadership meeting

Only 2% of CHROs are actually leading AI integration, even though two-thirds say it's a priority — that gap is our real risk.

Watch out for

Treating hiring, AI use/adoption, and workforce development as three separate problems instead of one interconnected capability system.
Accepting a candidate's or employee's first AI-generated output/response without pushing for iteration (accepting an answer within six prompts is a red flag).
Letting AI judge AI-enhanced work product ('work slop lasagna') instead of keeping a human in the review loop.

Fun fact · Alexander Alonso

He's served on the Secretary of Defense's Defense Business Board and racked up 400+ speaking engagements featured in USA Today, BBC, and CNN.

If this landed, see these

↳ Go deeperRedesigning Work in the Age of AI: Unlocking Productivity and Human PotentialTakes 'hire for demonstrated skill, not declared skill' further into a full job-redesign model for dividing tasks between people and AI.⇄ The counterpointFrom Typewriters to TikTok: The Power of Generational Wisdom within Modern LeadershipComplicates the claim that over-45 workers are AI's best users: generational patterns are real, but treating any age group as a fixed rule is the actual risk.✦ The unexpected oneCulture Redefined: Architecting a Skills-Driven, Continuous Feedback CultureSame demonstrated-over-declared logic shows up in performance management: real conversations and skills mapping beat documentation theater there too.