How to create an AI policy in a childcare program

At our session at the 2026 NCCA & NECPA National Leadership Conference in Memphis, we asked a room full of childcare directors and owners two questions.

First: how many of you are using AI on a weekly basis? Almost every hand went up.

Second: how many of you have a formal AI policy or training in place for your team? Silence.

A survey we ran with ECE leaders told the same story. Every single respondent who used AI was doing so regularly - but five out of six centers had never formally discussed AI use as a team. And when asked how important it was to receive specialized training on ethical AI use in childcare, the average response was a 4 out of 5. 

There’s a huge gap between how widely AI is already being used and how rarely it's being managed. People know they need guidance, but probably haven't had any formal training.

Your staff are almost certainly using AI right now and for good reason. AI can help reduce friction in the overwhelmingly busy environment of a childcare center. Here’s how many educators are already using it to streamline processes:

• Creating lesson plans
• Drafting newsletters and flyers
• Writing job descriptions 
• Responding to parent messages
• Answering inquiry calls and scheduling tours

Yet at the same time, without shared standards for how your team can use it well, you’re taking on risks you don't even know about. I’m here to help you how to address the risks while encouraging adoption.

How is AI being used in early childhood education right now?

Let's start with what's working, because there's a lot to feel good about.

When you think of AI, you probably think of tools like ChatGPT, Gemini, or Claude. These are called large language models (LLMs). These tools understand and generate language, and they're a lot more approachable than they sound. For those who are tech-averse, you genuinely cannot break anything. You can ask a question, see what comes back, and ask again if you don't like it.

The staffing crisis context matters here. ECE workers are among the most overworked, underpaid professionals in the country. So, as a result, AI can be a means for absorbing the tasks that pull them away from what they actually want to be doing: focusing on children.

In the centers we work with, that's showing up most in a few key areas. Lesson planning is one of the biggest. AI tools can generate activity ideas, adapt plans for different developmental levels, and help educators save 30 to 60 minutes of prep time per week. Our survey respondents backed this up. Curriculum development and content generation was the second most common area where staff were already using AI, right behind parent communication.

Parent communication is the other area where AI is making a real difference. Drafting newsletters, translating messages for multilingual families, responding to common questions - these are ways in which AI can do a lot of the heavy lifting. Tools like Famly's Sidekick help educators write polished, professional messages even in the middle of a busy day. And on the operations side, hours of administrative work can be reduced to minutes, from writing job descriptions to pulling together grant applications. For directors who are already stretched thin, this matters.

And, I’m hearing about more really creative and intentional use-cases. One director I spoke with shared a story that really stuck with me. She needed to run a professional development session on trauma-informed care, but knew that many of her teachers had experienced significant adversity themselves. She didn't want the training to cause additional harm. So, she asked ChatGPT to turn the content into a play script — eight scenes her team could act out together. They spent an afternoon laughing, learning, and covering heavy material in a way that felt safe and joyful. That kind of creative problem-solving is what AI can unlock when it's in thoughtful hands.

What are the risks of using AI in childcare?

Here's where it gets more complicated. And to be clear: these risks don't go away just because AI is useful.

The most immediate risk isn't AI itself. It's actually from a lack of standards around when and how to use it. When staff use AI independently, without any shared guidelines, you end up with inconsistent outputs, unreviewed content going directly to families, and exposure the director isn't even aware of. 

Think of it like every teacher using a different curriculum they found on Pinterest - some of it is great, some of it is wrong, and none of it is coordinated. Our survey found that lack of staff training and expertise was one of the top barriers keeping centers from expanding their AI use. That tracks. When there's no shared foundation, people either go rogue or avoid the tools entirely.

Risk 1: Protecting children’s data

First, and most importantly: children's data needs clear protection. Data privacy and security was the most cited barrier in our survey, and for good reason. Photos, names, ages, developmental details - this is among the most sensitive information that exists. When you upload data to a free AI tool, that information can be used to train the model further, meaning it may be stored, analyzed, or surfaced in ways you never intended. 

Even if the risk feels abstract, the exposure is real: a child's photo or personal details entering an external system is information you can't take back. As a director or owner, your AI training should make this explicit so the boundary is clear, not assumed.

Risk 2: A lack of differentiation

There's also what we'd call the mediocre middle problem. When everyone uses the same AI tools without customization, marketing materials, lesson plans, and family communications start to feel generic. 

You can already see this happening with enrollment flyers. Centers across the country are producing ChatGPT-generated materials that look and sound nearly identical to each other. Same illustration styles, same broad language, nothing that tells a prospective family what actually makes one program different from the next. That's a competitive problem. Families searching for childcare are trying to find the right fit for their child, and if your materials blend into the background, you make that harder for them and for yourself. 

What makes your program special are your values, your community, and your relationships. Each of these elements don’t come through in a default AI output. When creating flyers, websites, or writing documents, think about what AI drafts as a starting point, not a finished product. 

Risk 3: The illusion of productivity

Have you ever heard the term “efficiency theater?” The term refers to when teams mistake speed and volume as true progress. AI can create the illusion of productivity, generating a lot of content quickly without actually improving outcomes.

A lesson plan generated in two minutes that doesn't reflect your children's actual interests or developmental stage isn't saving time. You're just producing busy work. One person creating mediocre content with AI can actually put extra work on others who have to review and fix it. The better question is never "can AI do this?" It's "should AI do this, and will a human review it before it goes anywhere?"

Risk 4: Bias

And finally, but still importantly, bias is a less obvious risk. AI models are trained on existing data, and existing data reflects existing inequities. A widely reported case found that an AI resume-screening tool elevated candidates with male-coded hobbies and college majors, even with names removed. Bias doesn't require intent. It just requires biased training data. 

This matters acutely in ECE, a field that is more than 94% women and disproportionately women of color. If you use AI to support hiring decisions, look closely at the criteria it's applying and ask vendors directly how their model was trained.

What should an AI policy for a childcare center include?

You don't need a legal team or a tech background to do this. You just need to lead the conversation before the absence of one creates problems. Here's a practical framework to get you started.

Start by defining what AI actually means in your context. Many staff may not realize that tools they're already using in their daily life like Youtube, Grammarly, Canva, and even Instagram already leverage AI. Explain how these tools differ from LLMs. Starting there, with a shared definition, gets everyone working from the same place.

Clarify who can use AI, and for what. Not every task needs the same level of oversight. A useful way to think about it: 

• Lower-stakes uses like brainstorming activity ideas or drafting internal notes can proceed with minimal oversight. 
• Medium-stakes uses like newsletters, social posts, job descriptions, parent-facing communications should require a human review before anything goes out. 
• Higher-stakes uses, like anything involving children's personal information or hiring decisions, should involve director approval or be off-limits to AI entirely.

Make your data rules explicit and non-negotiable. Staff should know clearly: never upload photos of children to any AI tool. Never input a child's name, date of birth, health information, or identifying details. Never share confidential family information including custody arrangements, documentation records, financial details with any AI system. When using AI to support lesson planning, describe children in general developmental terms only. These rules help protect the families who trust you.

Set a "human review before it goes out" standard. Anything that leaves your center like a newsletter, a social post, a parent message - should have a real person review it first. AI-generated content needs to be edited for accuracy, tone, and your center's voice. Clearly define who is a part of the review process. 

Build in actual training. A policy without training is just a piece of paper. Even a 30-minute session is a meaningful start. Walk your team through what AI tools are, which ones are approved for which purposes, the data rules, how to write a good prompt, and when to loop in a director. Build this into onboarding so new staff start with the same foundation.

Bonus: Consider adding in discussions about into your quarterly staff meetings. Encourage staff to learn from each other’s uses of when/how AI supported their work.

Plan to revisit it. AI is moving fast. Build in a review point once a year to update your policy as tools and best practices evolve.

At this point, AI is almost certainly already in your center. It’s up to you to ensure whether they'll use it well. A clear policy doesn't have to be complicated. It just needs to be honest about the risks, specific about the rules, and grounded in the values that make your program worth trusting. Your families trust you with their children. That trust extends to every tool you use.