In January, I launched a web app with a half-formed idea: What if families could take pictures of real-world objects and get instant, kid-friendly audio stories in response?

At the heart of this experience was Miko, a meerkat mascot with a magnifying glass and an endless drive for curiosity. With the help of AI tools, I mocked up the first version of Miko, but I knew I needed help from real artists to bring the character to life. That’s where the High School of Art and Design came in.

Unlike most public schools in New York City, the High School of Art and Design is a portfolio school. That means every student who applies must also submit a portfolio of work ideally to align across one of eight disciplinary areas: fashion, graphic design, film, architecture, cartooning, illustration, photography, and animation. As a result, the students at this school are among the funkiest and more creative bunch you’ll meet in New York City. 

When I visited the school early that year, a poster caught my eye for a mascot and logo design competition for an upcoming event. When I asked Amanda Deebrah, the school’s internship coordinator, about opportunities to collaborate with interns at the school, she told me about Agency, one of the school’s work-based learning initiatives, where students are paired with local companies to take on semester-long design projects. I quickly submitted a proposal for my new company, MuseKat, and it was approved.

From March through May, ten students helped me research and brainstorm some early brand work for MuseKat. We met weekly. They shared design concepts or brand research. Meanwhile, I kept them abreast of the ever-shifting landscape of my app–in-progress and the tech industry overall. 

As a brand new founder with a newborn startup, I didn’t know exactly how to shape a semester-long program, but I drew upon my years of experience in conceptualizing the Etsy entrepreneurship pilot with the High School of Art and Design, building out micro-internships for FutureReadyNYC with Tech:NYC, and in co-creating tech pathways for the Summer Youth Employment Program from 2020 - 2022.

Over three months, we focused on:

1. Visual brand identity: Logos, fonts, and color palettes

2. Character design: Accessories, moods, and expressions

3. New product concepts: Merch, musical elements, and even wearables

Thanks to their research and mockups, I quickly leveled up on having my own design vocabulary to discuss some of the key concepts (including my color scheme and typography). Their work also highlighted some important trends for me in comparable kids’ learning apps, such as using more rounded and curved lines and opting for a color palette without a lot of black.

Here’s the logo and app concept I eventually landed on. While I ultimately used AI to help me create these version, you can see how much inspiration I drew from the student's concepts.

My favorite part was seeing the students draw concepts for their own meerkat creations, which helped me consider some of the other elements to include in Miko’s character card. I’ve since started to use other accessories with Miko, like headphones and binoculars. When, without prompting, a few students started to conceptualize merch ideas on their own, I expanded the assignment to ask them to consider how MuseKat might exist in more of a screen-free wearable device for kids. 

Here are some of the concepts they came up with:

While this wasn’t an internship about AI, it was impossible to ignore my prolific use of AI in the creation and iterative process of building out MuseKat. The students were well aware that I “vibe coded” the first version of the app, including the meerkat and the brand and logo concepts. 

As such a specialized school devoted to the craft of design, this introduced quite a few questions that we had to grapple with:

On our first day, I reminded students that they should only surface ideas that they felt comfortable with me using for MuseKat more broadly. It felt prudent to call out that I would never upload their original design concepts through AI to build iterations on my own. (Notably, there is no way of proving that I stuck to my word.)

With each iterative assignment, I invited students to either draw by hand or use AI to conceptualize different concepts. In the end, we did have a few students use AI to create both merch ideas and new character concepts. Other students felt more comfortable sticking solely to the craft of the original art.

Then, halfway through the semester, OpenAI released a new image model. I tested it over the weekend, and suddenly, Miko’s whole world expanded. The model understood the character in different poses, moods, and contexts. I had a logo, a sticker sheet, and a more refined visual style in under 48 hours. I was thrilled. Then ashamed.

What was I supposed to tell the students next week? To pretend like nothing had changed, or be honest about the truth? That in fact, AI had, in many ways, replaced my need for human iteration, at least for this phase of the work.

Sheepishly, I admitted that I had been able to use AI to update my entire design schema. But I also made it clear: It was their ideas (the sketches, the research, and their design instincts) that gave me the language to prompt more thoughtfully. The simplified lines and the playful poses were echoes of the work we had done together.

I shared the results and asked for their honest thoughts and feedback on that choice. I heard a mixed variety of opinions. Some thought that the quality still wasn’t strong enough to merit choosing AI over human designs. Others agreed, but also acknowledged the tough position I was in, as a solo, boot-strapped founder without a lot of resources. Maybe AI is okay as a design partner, in some circumstances.

For our final project for the program, we all worked together to brainstorm an array of new characters for the so-called "Mikoverse," including more fleshed out personalities and "areas of expertise." To demonstrate the creative potential of AI, I build custom GPTs trained on each character for our final day of class, and we listened to each character's "narration" of the same piece of art, then critiqued the readouts of the AI's output.

Some were good, others, not so much. So, we talked about why that was, and what I might do to better train the AI to generate a better output the next time. It was a perfect snapshot of what it means to create in an AI-native world: Ideate, prompt, evaluate, and refine.

For me, the decision to bring AI into the forefront of this creative immersion wasn’t just self-serving for moving faster with my company: It was also deeply personal.

When I graduated from journalism school in the midst of the 2008-2009 recession and first big wave of social media upending media, I couldn't get a job in the industry I’d trained to work in for four years. I felt betrayed by my school for not preparing me more for the reality of the labor market, one that would only grow increasingly hostile to journalists in the decade to come.

I got lucky, I told the students. I happened to stumble upon a job at the intersection of technology and consulting, an early pivot that completely changed my trajectory. But I also shared that I couldn’t, in good conscience, pretend that AI won’t profoundly change the way we creative and design art. If AI models can change so much in just two months that I needed to reshape how I used student design input, it’s impossible to imagine how different the industry will look in 3–7 years when they’re starting their careers.

These conversations weren't easy. We don’t have a playbook for how to introduce the messy middle of AI. But I’m so grateful to have had a chance to explore this in real time with students, and hear their honest take on where they each personally thought I should draw the line.

What these students created wasn’t just early brand work; it was a window into how the next generation is already thinking about the future of creativity, technology, and work.

Throughout our internship, they asked me many smart, hard questions. They brought creative ideas to life. And they challenged assumptions that many adults–myself included–hadn’t fully reckoned with.

In a world where the rate of change can feel overwhelming, it's certainly safer to wait for formal AI policies to be rolled out school-wide or district-wide. But that would be underestimating what students are capable of contributing. As this group demonstrated to me, they aren't just passively preparing for life in the "real world" – they are actively shaping it.

That's why I believe students deserve to be in the room when decisions are being made, to explore new tools with agency and context, to meet industry professionals who are willing to be honest about what’s changing, and what’s at stake.

After all, if we want AI to support human creativity, we need to start by trusting the next generation to lead us there.

Acknowledgements: Thank you to Sheen, Iggy, Neil, Titus, Amina, Claire, Dariana, Karen, Mariami, Ulla, Lucas, Michel, Leila, Penelope, Reagan, Hailey, Mark, Asher, Zheng, and Jordan, our fearless agency leader, for all of your creative collaborative energy. You were all right, you really were "in it with me" for these past few months.

Special thanks to teachers Molly Snyder Fink and Amanda Deebrah for trusting me to experiment with a few new creative frameworks with this design Agency initiative.