1. AI adoption is a change management challenge first — and a technology challenge second

One of Cameron’s most important points was that organizations often approach AI as a tooling problem when it’s really a change management problem.

When new AI capabilities arrive, the instinct is to immediately operationalize them:

• Which tools should we buy?
• How do we drive adoption?
• How do we measure usage?
• How do we get ROI?

But that skips a critical step: helping people adapt to an entirely new way of working.

At Medium, the team spent time observing, experimenting, and grounding their approach in company values before creating rigid expectations.

“There’s no playbook for how to do that.”

Instead of racing to implement rules, they focused on building trust, creating space for experimentation, and helping employees understand how AI could support their work.

For People leaders, that’s a useful reminder: the hardest part of AI transformation isn’t the technology. It’s helping humans navigate uncertainty.

2. The goal isn’t to use AI more. The goal is to use better judgment.

Most organizations are still focused on encouraging AI usage. Medium is increasingly focused on helping people determine when not to use AI.

As Cameron explained, access to powerful tools is no longer the differentiator. The differentiator is knowing where human expertise, creativity, and taste still matter most.

“The key differentiator is behaviorally having the judgment to be like, ‘I’m going to use AI over here. I am not going to use it over here.’”

That shift changes the conversation entirely.

Instead of asking:

• Are employees using AI?
• How many prompts are they writing?
• How much money are we spending?

The better questions become:

• Are people making better decisions?
• Are they producing higher-quality work?
• Are they creating more value?

Judgment is becoming one of the most important workplace skills of the AI era.

3. Don’t create AI policies in a vacuum

Many companies are rushing to publish AI guidelines from the top down.

Medium took a different approach.

When Cameron and the AI Leadership Council drafted their AI fluency expectations, they initially considered presenting them directly at an all-hands meeting.

Then they stopped.

They realized that approach contradicted the culture they were trying to create. Instead, they shared the draft with managers first and invited criticism.

“Please tear it apart if it seems super off.”

That decision transformed the rollout from an announcement into a conversation.

Managers became collaborators rather than messengers.

The lesson is simple: if AI is changing how work gets done, the people closest to the work should help shape the expectations.

4. AI fluency isn’t about mastering tools — it’s about staying curious

One of the most refreshing parts of the discussion was how Medium reframed AI fluency.

The first draft of their framework focused heavily on specific behaviors and tool usage.

Cameron eventually scrapped it.

Why?

Because nobody truly knows the best way for every employee to use AI.

Instead, the team anchored on behaviors:

• Stay curious.
• Experiment.
• Share what you learn.
• Help others improve.
• Build good judgment.

“We’re going to assume everyone here is very smart. We hired them for a reason.”

That mindset creates a very different employee experience than a compliance-driven approach.

Rather than forcing adoption, it encourages exploration. Rather than measuring activity, it develops capability.

And capability is ultimately what organizations need if they want to remain adaptable as AI continues to evolve.

5. AI resistance deserves curiosity, not dismissal

Cameron had one of the most thoughtful takes on AI skeptics we’ve heard on the show.

When employees raise concerns about AI, the response shouldn’t be to convince them they’re wrong.

Many of those concerns are valid.

People worry about:

• Quality
• Environmental impact
• Ethics
• Authenticity
• Long-term career implications

Ignoring those concerns only erodes trust.

“Any AI detractor has an extremely good reason for why they think it’s not a good idea.”

At the same time, Cameron believes People leaders have a responsibility to prepare employees for the reality of the job market.

As Head of People and Talent, he sees AI fluency appearing across job descriptions in nearly every function. That creates an important leadership responsibility:

Help employees engage with AI critically while also ensuring they develop skills that will remain relevant in the future.

6. If you want to understand AI adoption, stop measuring spend and start measuring sentiment

One of the most practical ideas from the episode was Medium’s AI Sentiment & Adoption Survey.

Rather than treating AI spend as the primary success metric, the team wanted to understand how employees actually felt about the technology.

They surveyed employees on two dimensions:

• Sentiment
• Impact

The results were fascinating.

More than 80% of employees reported experiencing some form of “AI brain fry” — the mental exhaustion that comes from keeping up with constant change.

At the same time, 70% said they would be disappointed if their AI tools disappeared.

“The pace of change alone, while staying on top of your normal job responsibilities, was the greatest cause of AI brain fry.”

The survey also uncovered an insight many companies are probably experiencing but not measuring:

The biggest source of friction isn’t creating work with AI.

It’s reviewing all the work AI creates downstream.

• Policies.
• Documents.
• Designs.
• Code.
• Meeting notes.
• Recommendations.

The challenge is ensuring the output is actually useful.

That’s why Cameron believes sentiment is a leading indicator while AI spend is merely a lagging one.

If employees feel overwhelmed, confused, or burned out, adoption metrics alone won’t tell the full story.