This work is grounded in the learning sciences, especially social design-based experimentation developed by Kris Gutiérrez and colleagues.
My framework draws on two key orientations:
Culturally sustaining pedagogies, which value and sustain students’ cultural and linguistic practices.
Consequential learning, which positions learners as capable of shaping the worlds they inhabit.
From this perspective, learning environments are designed spaces for collective inquiry, experimentation, and transformation. Learning is not separate from design; it happens through participation in changing the conditions people navigate.
My framework draws on two key orientations:
Culturally sustaining pedagogies, which value and sustain students’ cultural and linguistic practices.
Consequential learning, which positions learners as capable of shaping the worlds they inhabit.
From this perspective, learning environments are designed spaces for collective inquiry, experimentation, and transformation. Learning is not separate from design; it happens through participation in changing the conditions people navigate.
From this foundation, I propose a framework I call co-dreaming with algorithms.
The framework moves beyond simply using AI tools. It treats them as sites for questioning, imagining, and redesigning.
It is guided by three principles:
The framework moves beyond simply using AI tools. It treats them as sites for questioning, imagining, and redesigning.
It is guided by three principles:
- Creative and playful
- Intergenerational and interdisciplinary
- Future-oriented
The work presented here comes from data collected across a series of workshops with students and educators exploring AI, identity, and representation.
Each workshop begins with critical AI scholarship, then moves into hands-on exploration of facial recognition technologies. From there, participants “drag” these systems using tools like Midjourney, and end with reflection and design.
Each workshop begins with critical AI scholarship, then moves into hands-on exploration of facial recognition technologies. From there, participants “drag” these systems using tools like Midjourney, and end with reflection and design.
The workshop moves from critique, to experimentation, to imagination.
Participants first examine how facial recognition systems interpret them. They quickly notice discrepancies: misread ages, incorrect classifications, and results that do not align with how they understand themselves.
These errors are not random. They reflect broader patterns of bias, especially for people of color and marginalized communities. What begins as a technical activity quickly becomes a conversation about representation and power.
Participants first examine how facial recognition systems interpret them. They quickly notice discrepancies: misread ages, incorrect classifications, and results that do not align with how they understand themselves.
These errors are not random. They reflect broader patterns of bias, especially for people of color and marginalized communities. What begins as a technical activity quickly becomes a conversation about representation and power.
Participants then work with generative AI tools like Midjourney. The goal is not simply to create images, but to examine how these systems interpret and represent identity.
They ask: How can I alter my image in ways that disrupt or “trick” facial recognition systems?
Using images of themselves, participants generate altered versions that play with aesthetics, identity markers, and visual cues.
This process reveals a developing sociotechnical consciousness grounded in Chicanidad. For example, participants noticed that when they prompted “Chicano,” the system generated relatively white representations. But when they prompted “militant Chicano,” the images became noticeably darker.
This exposes how generative AI systems encode racialized associations and stereotypes. Participants are not passive users. They are probing, testing, questioning, and beginning to disrupt the system.
They ask: How can I alter my image in ways that disrupt or “trick” facial recognition systems?
Using images of themselves, participants generate altered versions that play with aesthetics, identity markers, and visual cues.
This process reveals a developing sociotechnical consciousness grounded in Chicanidad. For example, participants noticed that when they prompted “Chicano,” the system generated relatively white representations. But when they prompted “militant Chicano,” the images became noticeably darker.
This exposes how generative AI systems encode racialized associations and stereotypes. Participants are not passive users. They are probing, testing, questioning, and beginning to disrupt the system.
The workshop then shifts into structured reflection. Participants respond to prompts such as:
As a [self-identity], what I wish this technology would do is…
As an educator, what I wish this technology would do is…
In the future, I hope this tool will…
These prompts move the conversation from critique toward imagination and design.
Here, participants are no longer only naming what is wrong with AI systems. They are articulating what they want them to become.
In this example, the participant imagines an AI that recognizes them as they are, affirms queer identity, and does not force people into narrow or normative categories.
This is not just reflection. It is design. It is an articulation of values and of what a more just, responsive AI system might look like.
This is where speculative fabulation becomes visible: participants imagine futures that do not yet exist, but that are grounded in lived experience.
As a [self-identity], what I wish this technology would do is…
As an educator, what I wish this technology would do is…
In the future, I hope this tool will…
These prompts move the conversation from critique toward imagination and design.
Here, participants are no longer only naming what is wrong with AI systems. They are articulating what they want them to become.
In this example, the participant imagines an AI that recognizes them as they are, affirms queer identity, and does not force people into narrow or normative categories.
This is not just reflection. It is design. It is an articulation of values and of what a more just, responsive AI system might look like.
This is where speculative fabulation becomes visible: participants imagine futures that do not yet exist, but that are grounded in lived experience.
Finally, those reflections move into design through a prototype created in Playlab.
The bot reflects the desire we just heard: to create AI that is supportive, affirming, bilingual, and responsive to people’s lived experiences.
So this is not just a demo. It is a small prototype of a different kind of AI.
You can use the QR code to try the bot yourself.
The bot reflects the desire we just heard: to create AI that is supportive, affirming, bilingual, and responsive to people’s lived experiences.
So this is not just a demo. It is a small prototype of a different kind of AI.
You can use the QR code to try the bot yourself.
