Improving AI Literacy Engagement for Skeptical Media Arts Students
How do we get creative students who are worried about AI to think critically about AI without necessarily pushing them to adopt it?
COMPLETED:
METHODS:
Curriculum Design
Mixed-Methods Research
Learning Experience Design
Qualitative Analysis
Instructional Design
ROLE:
Director of Learning Experience · Curriculum & Product Design Lead (co-led with CTO)
IMPACT
Piloted with 70 students, ages 6–17, across 6 classes in 4 weeks
Outcomes led to the launch of a new nonprofit initiative: Creative Futures Collaborative
THE "MESS"
Students generally came in with a negative impression of AI, so bringing up AI use in creative work make them tune out
Supported by OpenAI's People-First AI Fund, OurWorlds and Media Arts Center San Diego piloted an AI literacy curriculum at MACSD’s 2026 summer camps. The initial approach that featured artists using AI in their work, was met by disengagement and opposition from students. 52.8% rated their worry about AI 4 or 5 out of 5, and one 16-year-old crossed out “Very Interested” on our survey scale and wrote “Worried.”
The goal was to land on an approach that felt considerate of student concerns and drove curiosity about AI technology, without necessarily increasing adoption.
To Do:
Build curiosity about AI without pushing adoption
Create a conversive atmosphere allowing students to air their feelings
Provide agency to students over whether AI is to be used
KEY INSIGHTS & DESIGN IMPLICATIONS
The biggest culprit for disengagement was bringing in AI topics amidst student anxiety about AI's negative impacts
This was both observed in the classroom and confirmed by surveys during pilot.
Students came in vocal and pessimistic about AI
Across pre-activity surveys, 52.8% of students rated their level of worry a 4 or 5 out of a 5pt scale. At the same time, students wanted a say: 16 of 36 wrote their own answers outside the options we gave them, on a summer camp survey.
Students want reassurance that AI's problems can be addressed.
Facilitators saw students get more attentive once AI’s resource use, biases, and limits were included in discussion. The August 4 cohort got the most of this content and reported the strongest engagement experience (qualitatively).
Student's lack AI literacy, but it is critical they have it.
When asked to rate their confidence in their ability to talk bout AI, 66.7% of students rated themselves in the bottom half of the scale. Asked what AI is good at, three 9–12-year-olds wrote, unprompted, “Absolutely nothing but being dumb,” “Stealing,” and “Nothing just being wrong wrong and wrong.”
Specifying where AI ought to be used made a difference.
Students engage when they see positive solutions, such as AI usage in wildfire mitigation and coral reef protection. Students are also protective of creative agency.
KEY INSIGHTS & DESIGN IMPLICATIONS
Iterating on the Pilot Curriculum
Instead of prioritizing AI examples and tools, we learned to center in on a discussion of what AI really is and and what it is and is not capable of.
Some students were not allowed to use AI or did not want to. We created AI-free versions of lessons that allowed students to view AI outputs for examples without prompting it themselves.
Adding examples of positive AI uses cases & solutions to AI resource consumption. Basically opening up the floor for students to think about: if the environmental challenges were solved, would AI be useful to you?
How AI is used matters.
Students engage when they see positive solutions, such as AI usage in wildfire mitigation and coral reef protection. Students are also protective of creative agency.
THE SOLUTION
A Suite of Educational Tools & Content that Prioritize a Critical Approach to AI Literacy
Classroom material and lesson planning prioritized meeting students where they were at. AI is a polarizing topic as it is and students were confused between what they were hearing online, from their parents, and observing in their communities. These approaches prioritized agency over adoption, learning gains from both content and interaction, and technical instruction about the inner workings of AI tools.
A new framework for lesson outlines that lead with questioning AI before seeing examples of it.
By the end of the pilot, every lesson followed one order: inform, then question, then (only then) learn to use AI. This spurred our creation of a product for having discussions with students and a set of examples of how people are trying to mitigate the consequences of AI.
Tradeoff: Class sessions were limited to 60 minutes, so increasing time spent on class discussion on AI instead of media arts practice decreased the content students engaged with, especially in a summer camp where students were not assigned homework.
A tool for students to participate in classroom AI use policy
Each class placed AI use cases on a shared quadrant: how good an idea is this, and how much pressure do you feel to use it? Students identified relationships between the two and discussed why using AI for one task feels better than another. With instructor input, the end product sparked an educator resource for producing a bottom-up AI use policy that apply to classrooms.
Tradeoff: There were still polarizing differences between some use cases of AI, such as using AI to generate images, so not all students were on board with every outcome. Also, doing this activity as a class creates more room for the louder voices to influence outcomes.
Lessons that use Student Craft as a Means of Evaluating Generative AI
Activities tested AI outputs in comparison to student work to stimulate critical thinking over what each does well. For students who’d already used AI, the value was a new vocabulary for describing what they already felt. The 9–12 cohort, all prior AI users, scored lower than ages 6–9 on “I learned something new” (2.29 vs. 3.47) but much higher on “I could explain AI to a friend” (4.14 vs. 2.18)
Tradeoff: For students skeptical about AI use, weak AI outputs were likely confirm the idea that “AI is dumb” before students saw more advanced examples of what it can actually do
IMPACT
Students Left Just as Skeptical, but Better Informed
Across the cohorts, the most common post-activity answer measuring AI attitude was “stayed the same” or “slightly more negative.” Only a few students moved more positive. But this was okay, since we set out to build informed views, not adoption, and the reasons students gave got more specific:
16 of 36 students wrote answers outside the options we gave, adding their own categories and correcting our labels.
In the 9–12 cohort, curiosity moved from how AI works to what it does to people. One student asked “how does it affect the others, the future, and world nature.”
Students named what they wanted next: copyright first, then water and electricity use, “AI psychosis,” and “slopification.”
HERE'S WHAT FOLLOWED
While we were focused on building AI literacy for students, we almost missed another user group: teachers.
Of the Media Arts Center educators we surveyed, 67% strongly agreed that students should learn to judge when AI is appropriate. However, only 33% felt prepared to help them do it, and none felt they could explain AI systems well enough to answer students’ questions.
IMPACT
Now, these lessons and tools are freely available for families and educators to use for their students at creativefuturescollaborative.org
REFLECTION
Having Agency over AI is a Critical in Education
AI is polarizing, especially within the media arts community. What was interesting to observe was that while initially completely opposed to talking about AI for the sake of it being AI, further iteration with learning material revealed that the fuel for this hatred was really a fear of having no control over their futures.
This is not something that can be solved with one product or even a suite of products. Designing for this "mess" requires coordinated, multi-disciplinary expertise and active iteration to keep up with the pace of AI development.