In conversations about AI adoption right now, “human-first” has become a kind of shorthand.

It’s used in internal emails, leadership statements, and panel conversations—often as reassurance. A way to say: we care about people, we value craft, we’re not about to automate everyone out of a job.

But as AI becomes more embedded in everyday business systems, “human-first” can’t just be a sentiment. It has to mean something operational. Something legible. Something teams can feel in how decisions are made, not just hear in how they’re announced.

This post is about what “human-first” actually means in an AI era: in business broadly, and in publishing specifically. It’s about codifying values into practice —so leaders can communicate clearly, teams can feel grounded, and organizations can move forward without betraying the people who make the work.

“Human-First” Is Not Anti-Technology

Being human-first does not mean being anti-AI. It doesn’t mean refusing to engage with new tools or pretending technological change isn’t happening. In fact, many of the strongest human-first frameworks come from organizations that actively work with advanced technology.

A human-first approach can sometimes mean dealing head-on with intense organizational opposition to AI adoption. A lot of those concerns are real, and assuaging them will be the first order of business.

Most AI tools and systems are marketed as “human-first.” But that will only be true of your AI policy implementation if you listen to and educate the team at the outset and all the way through tool adoption and rollout. Selecting representatives to help speak for the different business divisions will give you grounded feedback on what’s working, what’s not, and what’s freaking people out.

This active listening will help you avoid software sprawl while easing the delicate tool rollout and adoption phase. It will ensure you’re meeting staff where they are in their understanding of AI and bringing everyone along together. And that’s how you equip people to continue to use AI for the drudgery and increase time spent strategizing, rather than leaning on tools they may not understand. That dilutes your AI strategy’s effect. In other words, human-first isn’t about saying no to technology. It’s about deciding how and where technology belongs.

Why This Conversation Feels So Fraught in Publishing

Publishing’s anxiety around AI isn’t irrational.

This industry has lived through decades of consolidation, layoffs, shrinking margins, and “efficiency” initiatives that often landed hardest on the people doing the work. Against that backdrop, any new technology can feel less like an opportunity and more like a threat—especially when it’s framed in terms of speed, automation, or cost reduction.

Publishing also trades in something uniquely fragile: trust.

→ Trust between authors and editors.
→ Trust between agents and publishers.
→ Trust between teams asked to do more with less.

So when leaders talk about AI without grounding it in values, people fill in the gaps themselves. And the stories they imagine are rarely generous.

Being human-first is about closing that gap before fear does the talking.

What “Human-First” in an AI Era Looks Like in Practice

A human-first AI approach starts with clarity on decisions about how AI will be thought about, implemented, and messaged at your organization. 

Here are some of the principles we see in organizations—inside and outside publishing—that are handling this transition well.

1. Craft Stays Human; Drudgery Goes to AI

This distinction matters enormously in publishing.

Human-first organizations are explicit about where AI does not belong: in creative authorship, editorial judgment, taste-making, and relationship-building. These are the core human competencies publishing depends on.

At the same time, they’re honest about how much repetitive, administrative, and data-heavy work has quietly expanded across every role.

Human-centered design in AI focuses on relieving cognitive load—freeing people to do the work that actually requires their creativity and deep strategic thinking.

In publishing, that can mean:

• using AI to build and update production schedules
• summarizing author questionnaires
• flagging metadata inconsistencies
• drafting internal reports or campaign summaries

None of this replaces judgment. It protects it.

2. Humans Remain Accountable

Another hallmark of human-first AI is clear accountability.

AI systems can suggest, surface, summarize, and analyze—but decisions remain human-led. When something goes wrong, responsibility doesn’t disappear into “the algorithm.”

Human-first AI frameworks keep accountability visible and assignable, even as systems become more complex.

For publishing leaders, this is critical. It reassures teams that AI won’t become a shield for opaque decisions or a way to avoid hard conversations. The buck has to stop somewhere; this encourages ownership when things go wrong and leaders who have the institutional knowledge and AI training to avoid repeating the mistake again. The alternative is relying blindly on systems we know aren’t perfect and sometimes hallucinate, all in the name of plausible deniability.

3. Values Are Codified, Not Implied

Human-first organizations don’t rely on vibes.

They write things down. Then, they amend those documents as needed. They plan for this to be a long-term, iterative process rather than a one-and-done statement.

They articulate boundaries: what’s allowed, what’s off-limits, where review is required, and how decisions will be revisited. These aren’t static rules—they’re living guidelines that evolve as tools change.

Codifying values does two things:

• It gives teams clarity.
• It gives leaders a shared language for communicating change.

Without that, even well-intentioned messages can land as hollow reassurance.

4. Learning Is Treated as Essential

One of the most overlooked aspects of being human-first is education. Publishing has long taken an “apprenticeship” approach to hiring new members. Your boss teaches you what their boss taught them. No strategic or tactical training from the organization itself. Better hope you have a good boss (and they did too). 

Organizations that leave teams to “figure it out on their own” are creating anxiety in those teams.

Human-first companies invest in shared literacy: what AI is, where it’s already embedded, what risks actually matter, and how decisions are made.

In publishing, learning reduces fear because it replaces mystery with context. It empowers more team members to understand the work they’re doing deeply–so they can be someone who spots the next great time-saving tool or more efficient workflow. In a setting where trainings are regularly held and enthusiastically designed, they’ll feel comfortable tossing out their next great idea.

Communicating “Human-First” to Your Team

This may be the hardest part.

Teams don’t need leaders to promise that nothing will change. They know that’s not true.

What they need is honesty:

• about what is changing
• about what is not
• about what leadership is still figuring out

Human-first communication sounds like:

• “Here’s where we believe human judgment is non-negotiable.”
• “Here’s where we think AI can relieve pressure.”
• “Here’s what we won’t do.”
• “Here’s how we’ll revisit these decisions together.”

That kind of clarity builds trust even when answers are incomplete.

Why This Matters Now

AI is moving quickly. What publishing needs is leadership willing to articulate values in a way that can survive real-world pressure.

A human-first approach doesn’t slow innovation. It makes it sustainable. It ensures that as tools evolve, the people doing the work don’t feel erased, sidelined, or lied to.

Publishing has always been a human industry. That doesn’t change.

What changes is how intentional we are about protecting that humanity as systems around us evolve.