Are HR and L&D still hesitating on AI, or are they ready to take the lead?

Season 3 Episode 8 of our AI for the Average Joe Podcast Show features HR leader Michelle Hartley, who shares what many senior HR and learning and development professionals quietly admit: lots of dabbling, plenty of fear, and too much “wait and see – while business stakeholders are already using AI (sometimes well, sometimes badly).

HR and L&D, it’s time to stop lurking. Because that gap is your leadership opportunity.

HR, L&D and AI: the uncomfortable truth

Your managers and suppliers are already running HR queries, drafting investigations, and writing policies with AI. Sometimes they miss something critical – like the legal requirement for an appeals process – and you pick up the pieces.

That isn’t a case for banning AI. It’s a case for setting guardrails, standards and ownership.

This is where HR and L&D must transition from gatekeepers to governors.

Treat AI like a junior colleague, not a magic oracle

AI is superb at first drafts, options, data wrangling, and accelerating digital learning asset creation. It is poor at context, nuance, and judgement. Think of it as a capable junior: brief it clearly, review its work, send it back when it’s off, and make the final call.

What that looks like in practice:

Critical thinking as default. Ask “Where did this come from?” “What’s missing?” “What would change my mind?”
Source discipline. Tell your model which sources to lean on (policies, case law, internal frameworks). Don’t let it improvise.
RAG over ragged drafts. Use retrieval-augmented generation (even in lightweight forms) to ground responses in your documents, not the open web.
Own the outcome. If a candidate used AI to craft an application, assess evidence and experience in interview. Don’t police tools—validate capability.

HR and L&D to wear two hats: adoption and education

Senior people leaders now wear two hats.

Adopter: use AI to accelerate HR operations, workplace learning design, and people analytics.

Educator: enable managers to use AI safely and effectively. If you don’t fill this vacuum, shadow AI will.

Data & privacy

Define no-go data zones, redaction rules, and approved tools. Train teams on what can/can’t enter prompts.

Bias & fairness

Run regular audits on AI-supported decisions (recruitment, performance, reward). Require human sign-off for anything high-stakes.

Transparency & accountability

Be explicit: where AI is used, how it’s reviewed, and who is accountable (hint: humans).

Legal hygiene

Bake statutory steps (e.g., appeal rights) into your process templates and your AI system prompts. Never outsource compliance to a model.

Shift the talent playbook: assessment over authorship

If a better prompt can game your selection processes, the problem isn’t AI – it’s your assessment design. Move from evaluating polish to validating proof. Use work samples, job simulations, and live problem-solving. That’s how you future-proof selection in an AI-enabled market.

Personalise the machine; standardise the method

One of Michelle’s most practical tips: configure your model to your voice and standards. Use custom instructions so outputs reflect your organisation’s tone, risk posture and inclusivity guidelines. Then standardise how teams brief, review and store outputs. Personalisation for productivity; standardisation for quality.

A pragmatic 30-60-90 to get out of “wait and see”

It’s time to tip momentum.

Here’s a pragmatic 30-60-90-day plan to shift from “Should we use AI?” to “How do we use it well?”

First 30 days:

Run an AI readiness scan across HR and L&D: use cases, data, risks, quick wins.
Publish a one-page AI usage policy (what’s OK, what’s not, and when to escalate).
Launch an AI quick wins sprint (e.g., policy drafting assistant, learning outline generator).

Days 31–60:

Pilot grounded HR copilot use (policies, case templates, behavioural frameworks).
Redesign two selection steps to emphasise evidence and live assessment.
Train managers in AI literacy: briefing prompts, validating outputs, maintaining audit trails.

Days 61–90:

Formalise governance (review board, bias audits, incident response).
Measure impact (cycle time, error rate, employee experience).
Scale to adjacent processes (onboarding, upskilling pathways, learning and development analytics).
The human advantage (and why it still matters)

Remember, AI is process at scale. Your edge is meaning, ethics, and context. The future skills that matter – judgement, systems thinking, coaching, and design – become the frame that makes AI valuable and safe.

Lead with those, and you won’t be displaced by automation; you’ll be amplified by it.

And if you need some help, you know where to find us.

Blog post co created with ChatGPT 06.10.2025