Over the past decade, registration forms, staff applications, and customer relationship management have increasingly moved online. As technology advances, it provides numerous opportunities to optimize summer camp operations. But where does artificial intelligence (AI), specifically ChatGPT, fit in?

Introducing ChatGPT

ChatGPT, developed by OpenAI, is a chatbot designed to answer questions and assist with repetitive tasks. For example, if you’re running a day camp for five weeks, you might ask ChatGPT for theme suggestions and receive ideas like Superhero Academy, Around the World, and Time Travelers, among others.

Beyond a Basic Google Search

Generative AI, like ChatGPT, is more advanced than a simple internet search. Instead of just retrieving links, it generates responses by analyzing a vast array of learned data and algorithms. It simulates conversation, synthesizing information that it predicts will be relevant, based on its training.

Understanding ChatGPT

The more you understand how ChatGPT works and the more you have conversations with it, the more refined the answers get. But we’re jumping a little ahead — we’ll get back to that later.

Using generative artificial intelligence like ChatGPT can be scary. Not only is it a new skill that truly takes time and learning to develop, but it also comes with a litany of moral, ethical, and legal questions. If it’s generating answers for you, who owns that content? Is it okay to claim ownership of what you work with it to create? What data or information is safe to give it? Where’s the line?

Talk to a handful of different people, and you’ll get a handful of different answers. And that’s OK. When it comes to generative AI, some things we can know — like how data is protected, who owns content, and how to get the best answers — and there are some things we have to decide for ourselves — like how we present cocreated information, what information we’re OK putting in, and how we’re going to use it.

Responsible Use of Generative AI

To start, it’s a good idea to do your own research. Make a list of questions that might impact how you use ChatGPT (or other generative AI), and get the answers to those questions. Then keep looking. Read a few more articles, and get a good idea of what the pros, cons, and legalities of usage are. The first step in responsibly using generative AI (and, really, responsibly using anything) is having a foundational understanding of how it works. From there, you can decide how you want to use it.

Build Policies on Your Values

Each organization’s approach to AI will differ based on its values. PHOS Creative, a digital marketing agency, publicly shares its AI policy. The agency encourages considering questions such as (West, 2023):

• What impact will AI have on my character and habits?
• Will it outsource tasks that should ethically remain human?
• What are the broader implications of my AI usage?

Answering these questions with your team can ensure that AI integration aligns with your values, reducing ambiguity and optimizing its use.

Consider Data Security

Exact data security varies depending on what generative AI you’re using. ChatGPT doesn’t store personal data and has regular updates for security purposes. All of this is important to understand before inputting data, whether that’s through conversations or direct uploads. (Yes, you can upload into ChatGPT — I promise more’s coming!)

At the end of the day, you have to decide what you’re comfortable uploading. I probably wouldn’t risk uploading protected health information. But I would feel good about uploading a list of names along with their activity preferences to help build out a camper or staff schedule. Again, that’s a line that I’ve set for myself and encourage you to do the same.

Recognize Biases

While working with ChatGPT to create an announcement for my assistant director, it assigned pronouns based on her name (Chesney), but it didn’t get that assumption right. That’s just the very tip of the iceberg of biases built into AI trained by biased data and biased humans. The very nature of generative AI is based on it making assumptions — and assumptions can be dangerous.

According to the University of Kansas’s Center for Teaching Excellence (2024):

ChatGPT, Microsoft Copilot, and other AI chatbots are trained on an enormous amount of publicly available online text. As a result, they have the same biases as those sources and the society that produced them. These include:

• They are dominated by a white, male perspective.
• They are highly influenced by American culture, American capitalism, and the English language.
• They can and will invent sources, people, and events. (The industry term is “hallucinate.”)
• They can generate offensive responses, although filtering has reduced the likelihood that will happen.
• Similarly, AI detectors are more likely to flag the work of writers for whom English is not a first language.

Definitely not perfect, especially for an industry that, generally, is making great strides toward justice, equity, diversity, and inclusion. This doesn’t mean we can’t use it, though. Just like we learn, grow, and provide space for ourselves and others to do that, we can use generative AI with deliberate awareness. That means acknowledging that these biases exist and looking for them in the answers we’re getting. It also means being humbly skeptical of the answers we get. AI isn’t always right.

When I get an answer that’s questionable or doesn’t sit right, I often follow up with additional questions, do outside research, and/or call ChatGPT in. For example, an answer it gave me recently referred to something as a “pow wow,” and I responded that I try to stay away from that and gave the reason why. When we are aware and speak up, we can help generative AI improve.

Read and Refine

Whether you’re looking for bias or something else, I can’t overstate the importance of reading and refining what you get from ChatGPT. Don’t accept the answers as perfect and done. Check it and find ways to make it yours. Insert more of your tone and voice to make it sound more human and better reflect you.

Practical AI Applications in Camps

Spreadsheet Assistance

I know — maybe not the “sexiest” one to start with, but spreadsheet assistance is one of the things I’ve found most handy so far. When there are things I wish a spreadsheet would do but I’m not sure how to do it or if it’s even possible, ChatGPT is the first place I go. I’ve done this two ways: explaining the issue and asking how to do it, and directly uploading the sheet and asking how to do it. I’ve found direct uploads more effective, but depending on what you’re working with, asking how to do it may better protect data.

Here’s an example: I had a spreadsheet with all of our staff and the weeks they were working. It looked something (not exactly) like:

Name	Allison
Session 1	Working
Session 2	Not Working
Session 3	Available
Session 4	Working
Session 5	Working

 

I wanted to send them an email for the weeks I had them down to work, but making a manual list of who was working when was going to be an annoying and timely task. I uploaded the sheet to ChatGPT and asked for a way for info to go in the next column that would help make a list of what weeks people were working, and it wrote a script for me. It told me step by step how to enter the script, and when I ran it, I had the list I needed. From there, it was easy to do a mail merge and email out when everyone was working. What probably would have taken me an hour or more took me just five minutes.

This is just one example, but it’s written many other formulas and scripts for me. If you can think of it with a spreadsheet, it might just be possible, and ChatGPT’s help can bring it to life.

Quick Organization/Formatting

Ever had someone send you something in a totally unusable format? Often, ChatGPT can help. For example, someone sent me a paragraph-style list of email addresses to import into our content management system (CMS), and our CMS only accepts CSV files. ChatGPT quickly made a CSV list out of the data for me.

Communication Proofreading and Clarity

Especially if you’re not the most confident writer, ChatGPT can be a great tool for improving communication. I’ll often copy and paste an email that I’m writing into ChatGPT and ask it to check for spelling and grammatical errors. Sometimes, I’ll even just say, “Make this better,” and it does (most of the time).

Going a step further, I’ll do the same with information I’m going to send out to families or staff, and I’ll ask ChatGPT what questions it might have after reading my email. Then, I go back in and clarify or add information before it goes out, making my communication clearer and more efficient.

Social Media

AI can also be a powerful tool for managing your camp’s social media. It can help generate content ideas, draft posts, and even analyze the performance of different types of content. For example, you might use ChatGPT to brainstorm engaging questions or poll topics that encourage interaction from your followers. Additionally, AI can suggest improvements to the wording of your posts to increase engagement or help tailor content to different platforms, ensuring that your messages are optimally received by each unique audience.

A pro tip that is useful across many applications of ChatGPT, and one that I use frequently for social media, involves the “Act as . . .” prompt. Before I ask for help with captions, I tell ChatGPT to “Act as a social media expert.” This simple directive yields responses that incorporate more emojis and are generally more engaging. The specificity of your prompts plays a crucial role in the quality of the responses you receive.

Acitivity Buildout and Write-ups

Have you ever had a nugget of an idea but needed to flesh it out more? ChatGPT is great for ideation. For example, you can tell it something like:

I’m thinking about doing an activity for 300 campers and 65 staff. We have one hour, limited space, and all basic sports and arts and crafts supplies. The theme of the week is pirates. So far, I’m thinking I want it to be a station game with 10 stations where groups of campers (10ish per group) can complete some pirate-related tasks that will help them with a larger mission that covers the whole activity. I want a cool kickoff and ending. Can you help?

When I did that, I got a cool theme, cool kickoff, and the following stations:

• Cannonball Chaos — Using black balloons (cannonballs), groups must knock down a set of targets.
• Swab the Deck — Relay race where campers must “clean” a section of the field with brooms or mops.
• Captain’s Quarters — Solve a puzzle to unlock a chest containing the next clue.
• Walk the Plank — A balance beam activity over a small pool (could be just a blue tarp with obstacles).
• Treasure Map Tangles — Teams piece together a map using scraps found at the station.
• Sea Monster Dodge — Navigate through a series of obstacles while avoiding “sea monsters” (staff with pool noodles).
• Message in a Bottle — Retrieve a message inside a water-filled bucket that can only be accessed using provided tools.
• Flag Hoisting — Teams work together to raise a flag using a makeshift pulley system.
• Blindfolded Rowing — Team members must guide a blindfolded “rower” through a course.
• X Marks the Spot — Dig for a token in a sandbox that will be part of the final treasure.

Not bad for taking my idea and running with it! If I were fully planning this out, I could chat back what I liked, what I didn’t like, and what I wanted to build out more. It’s great for creating together.

You can do something similar with rules for games or staff training sessions.

Menus and Special Diets

ChatGPT can provide a great draft when you want to redo your menu! You can specify number of people, number of meals, special diets and allergies, budget, preferences, and more, and it can help you create a great menu. Again, the more you work with it, the better the response will be.

Embracing AI Responsibility

If you’re not convinced, give it a try yourself. I’d love to hear what values-based generative AI policies you come up with and how you’re using AI.

 

The University of Kansas. (2024). Helping students understand the biases in generative AI. Center for Teaching Excellence. cte.ku.edu/addressing-bias-ai

West, B. (2023, August 15). A value-driven generative AI policy for PHOS. phoscreative.com/articles/ai-policy

Author’s Note: The insights in this article were enhanced with input from ChatGPT, OpenAI’s language model, which generates text based on patterns and data from its training.

Allison Krabill is the executive director for The Summer Camp Society and a former Y camp executive. She can be reached at allison@thesummercampsociety.com.