NARST 2026 · April 21

Community, Transparency & Tinkering for Just Futures

Lessons Learned from the Critical AI Literacy Institute (CALI)

Luke Waltzer, Laurie Hurson, Zach Muhlbauer, and Sule Aksoy

CUNY Graduate Center, New York City, NY, USA

Luke Waltzer

The Critical AI Literacy Institute: Asserting and Preserving Scholarly Agency in the Age of AI

CALI Director

The Teaching and Learning Center, CUNY Graduate Center

Luke Waltzer

About the Teaching and Learning Center

Educational development for graduate students
Research and advocacy
Open education and educational technology

Luke Waltzer

About the Critical AI Literacy Institute (CALI)

Origins
Elements
- Faculty Development
- Research
- Research & Development
- Advocacy

Luke Waltzer

Grounding Scholarship

Critical University Studies
Critical Ed Tech and Infrastructure Studies
DH, Science Education, Educational Development

Luke Waltzer

Defining Critical AI Literacy

Critical AI Literacies
Comprehensive AI Literacies

“Critical Artificial Intelligence Literacies (CAILs) is the collection of ways of thinking about and relating to so-called artificial intelligence (AI) that rejects dominant frames presented by the technology industry, by naive computationalism, and by dehumanising ideologies. Instead, CAILs centre human cognition and uphold the integrity of academic research and education.”
Olivia Guest, Marcela Suarez, Iris van Rooij, “Towards Critical Artificial Intelligence Literacies”

“This principle applies to all stakeholders in the educational ecosystem: it is the student’s agency to question, create with, or consciously decide not to use AI based on the task; it is the teacher’s agency to design learning experiences that align with instructional values, rather than ceding pedagogical control to a tool. True literacy involves teaching about agency itself, framing technology not as an inevitability to be adopted, but as a choice to be made. This requires a deep commitment to critical thinking and a robust understanding of epistemology.”
Tadimalla, Cary, Hull, et al, “Comprehensive AI Literacy: The Case for Centering Human Agency”

Luke Waltzer

The CUNY Context

Large, federated system
Heavy turnover at central office
Emphasis on career-connected learning
Minoritized population

Map of CUNY campuses across the five boroughs

Luke Waltzer

Our Goals

Reasoned Adoption Informed Refusal
Communities of Practice
Research, Tinker, Advocate

Luke Waltzer

Funding

Google.org
3 years / 1m
Second grant: Empire AI Initiative
Supportive relationship

Luke Waltzer

CALI's Strategy

Emphasize agency
Identify small teaching moments
Ground interventions in critical inquiry
Joy

Laurie Hurson

Fostering Critical AI Literacy as Collective World-Building: Curricular Models for Teaching With/About Generative AI

CALI Curricular Lead & Assistant Director of Open Education

The Teaching and Learning Center, CUNY Graduate Center

Laurie Hurson

Overview

CALI Curriculum
Faculty Interventions
Fostering Critical AI Literacy as Collective World-Building

Laurie Hurson

The Critical AI Literacy Institute

Faculty Development Program
Disciplinary Questions
Develop Critical Stance / Criticality (Sano-Franchini, 2026)
Reflective, collaborative, interdisciplinary

CALI faculty questions spanning Disciplinary Knowledge, Critical Considerations, Agency, and Pedagogies

Laurie Hurson

Summer Institute

Focus: Teaching with, about, against AI
Methods: Disciplinary working groups, affinity discussions, tinkering, mapping (Futch & Fine, 2012)
Goals: Develop interventions; foster a community of practice; model critical inquiry (Lave & Wenger, 1991; Haraway, 1988; Harding, 1993)

Laurie Hurson

Faculty Interventions

Research and Information in the Digital Age — Sarah Cohn, City College
Integrated Reading and Composition — Krystyna Michael, Hostos Community College
Introduction to History and Literature — Martha Nadell, Brooklyn College
Statistical Methods in Earth and Atmospheric Sciences — Spencer Hill, City College

Laurie Hurson

Research and Information in the Digital Age

Approach

8-class module
GenAI output assessment and model comparison
In-class survey, discussions, group work, reflective writing

Syllabus

Introduction to LLMs
Evaluating LLM output
Labor, environment, & ethics
Plagiarism, cheating, & academic integrity
Cognition
Creativity
Alternative models & resistance
Futures

Laurie Hurson

Integrated Reading and Composition

Creativity and labor
Students’ lived experiences and expertise
Reflection on value of writing

Laurie Hurson

Introduction to History and Literature

Readings on genAI and how it works
Collaborative AI class policy
Custom bots / model for brainstorming, genre writing

Martha Nadell's brainstormer custom bot interface

Laurie Hurson

Statistical Methods in Earth and Atmospheric Sciences

“What are all the ways that genAI influences my learning?”
Discussion of genAI usage and student learning
“A/B” test and Claude Code

“I entered the program as something of an AI ‘booster,’…I came to a much more nuanced stance…through CALI I came to find that AI presents enormous risks to student learning—alongside its potential upsides—to the extent that it feels like one of the single greatest challenges facing higher education in coming years: how do we create norms, tools, and whatever else we need for AI to become predominantly a tool for accelerating student learning rather than a crutch that supplants it?”
Spencer Hill, City College · CALI faculty reflection

Laurie Hurson

Interventions for Teaching Critical AI Literacy

Frame AI as sociotechnical system (cultural, political, economic, material, ecological)
In addition to skills, consider values, power, futures
Allow for refusal and/or adoption of GenAI tools
Leverage community and collaboration
Explore knowledge production practices

Laurie Hurson

Critical AI Literacy & Student Agency

Reflect on value of labor, friction in learning process
Widen lens of analysis, future is not determined (Cottom, 2026)
Students can refuse, ask questions, decide if/how tools are used
Give students control over the tools that they use

Laurie Hurson

Critical AI Literacy, Agency, & Collective World-building

“The classroom remains the most radical space of possibility in the academy.”
hooks, 1994

Onto-epistemology — “knowing is a material practice of engagement as part of the world in its differential becoming”
Barad, 2007

“There is good reason to believe vision is better from below the brilliant space platforms of the powerful… ‘subjugated’ standpoints are preferred because they seem to promise more adequate, sustained, objective, transforming accounts of the world.”
Haraway, 1988

Resist inevitability narratives
Examine across scale and contexts
Interrogate systems of power
Build solidarity, resistance

Zach Muhlbauer

Tinkering as Critical AI Literacy: Teaching AI Infrastructure through Breakdown and Reconfiguration

CALI Technical Lead and PhD Candidate in English

The Teaching and Learning Center, CUNY Graduate Center

Zach Muhlbauer

Tinkering as Critical AI Literacy

Design proceeds through recombination, not invention from scratch
Ends shift as available materials reveal new constraints and possibilities
Faculty work this way with models, prompts, datasets, and course materials
Tinkering turns those assembled parts into local, course-specific tools

Zach Muhlbauer

Tinkering in Pedagogy

Raising questions each iteration opens with a fresh probe
Reflection reflection-in-action at small scale (Schön)
Hands-on knowing tacit knowledge gained only by doing and failing
Undirected exploration bottom-up play converges into focus (Resnick)
Problem framing constructing the problem from puzzling materials
Personal toolbox expanding building blocks through playful use
New applications matching tech first to uses “new for the student”
Design decisions small-scale choices train judgment under incomplete knowledge

Zach Muhlbauer

Infrastructure and Instructional Design

Identifying pedagogical challenges faculty face in their classes
Translating disciplinary methods into technical control over AI tools
Prototyping domain-specific AI tools for classroom application
Pilots: Spanish (Hunter, Baruch) · English, History, First-Year Writing (CCNY)

Zach Muhlbauer

Teaching Infrastructure through Breakdown

Where does the AI model end and the user interface begin?
Consumer chat platforms obfuscate the difference (and its contingency)
Susan Leigh Star: infrastructure becomes visible on breakdown

Zach Muhlbauer

Hyperparameters as Instructional Controls

Temperature — shapes the probability distribution of an output
Low: steady, predictable
High: loose, surprising
At 2.0, coherence breaks

Zach Muhlbauer

AmigAI: Toward Breakdown and Remediation

Conversational language-learning tool running on Gemma-3-27b
Piloted in heritage and non-native Spanish classes at Hunter & Baruch
Flattens regional dialects into a dominant Spanish register
Reveals absence of low-resource languages in pretraining data
Underscores value of fine-tuning AI on low-resource languages

AmigAI-Template chat in Open WebUI: a Spanish greeting exchange where the model introduces itself as Ana and offers to discuss migration, health, or gentrification in New York

Zach Muhlbauer

Tinkering with System Prompts

Revisions tracked across participating campuses
Settings shift with pedagogical goals
Faculty test, compare, revise, repeat
AI model configuration as instructional design

Baruch. Prompt rewrites as lesson plans, each anchored to one assignment.

Hunter. Prompt and parameters tuned together: temperature 0.7 → 0, companion to scripted role.

Brooklyn. Six Spaces in parallel, one persona per assignment. Temperature 0 → 2; five models tested within a single Space.

Zach Muhlbauer

The CUNY AI Lab Sandbox

Self-hosted instance of Open WebUI
Open-weight language models from small to large in a single interface
Only model providers with zero data retention and training turned off
Compare models side-by-side, write system prompts, adjust parameters, export transcripts
Model cards with preset system prompt tied to knowledge base, agentic skills, tool calling

CUNY AI Lab Sandbox interface anatomy: AmigAI-Template preset model card with new chat, model selector, model card info, and prompt input — Anatomy: **(1)** new chat · **(2)** model selector · **(3)** model card info · **(4)** prompt input

Side-by-side comparison in the Sandbox: AmigAI-Template (left) and Google Gemma 3 27B base (right) responding to the same Spanish greeting prompt — Anatomy: **(1)** new chat · **(2)** model selector · **(3)** model card info · **(4)** prompt input

Şule Aksoy

Beyond the Black Box: Resisting AI Inevitability Rhetoric and Implications for Science Education

Assistant Professor of Science Education, SUNY Brockport

CALI Research & Evaluation, CUNY Graduate Center

Şule Aksoy

AI inevitability rhetoric as a knowledge claim deserving scientific scrutiny

AI as a colonial project; imperial practices of domination of people, land, culture, education, and politics (Hao, 2025)
AI as an embodied and material technology (Crawford, 2021)
AI as an automated deployment of violence; genocide-epistemicide-ecocide (Ricaurte, 2019)

Şule Aksoy

Conceptual Background

Limited critique and interdisciplinary approaches in science education (i.e., Avraamidou, 2024; Heeg & Avraamidou, 2023; Li et al., 2023)
Call for critical transdisciplinary, transnational, and ecological (TTE) approach to science and science education (i.e., Phillip et al., 2025; Morales-Doyle, 2023)
Critical consciousness and social justice in science education (i.e., Freire, 2000; Morales-Doyle, 2019; Bang et al., 2012)

Şule Aksoy

CALI as evidence

Faculty reflections on the inevitability
Written reflections from online meetings focusing on knowledge production, agency, and pedagogy
A qualitative case study

Şule Aksoy

Data analysis

Initial coding of faculty reflections
Comparing initial codes
Categorizing emerging themes
Memo writing
Sharing the synthesis with the faculty to reflect on

Şule Aksoy

Critique of Techno-determinism

“… presenting it as inevitable leads to pressure of a sort, to use it, to not critically interrogate it, to accept it as neutral or unbiased and not reflective of those who create it and the society it reflects.”
CALI faculty reflection · English

“Not incorporating AI into my courses, … feels to me irresponsible, even while recognizing other legitimate concerns one may have about the whole generative AI enterprise. How to do this effectively given the many binding constraints … seems to me the billion-dollar questions. What are the incremental things one can do?…”
CALI faculty reflection · Earth Science

Şule Aksoy

Threat to Agency

“… I agree that the interests behind genAI are clearly sending us the message that if it’s not unavoidable now, it will be soon. Get on board or get left behind.”
CALI faculty reflection · Library Sciences

“I’m interested in pushing back on the narrative of AI inevitability – that there is only one outcome for this technology… I’m not sure what institutional support in counteracting this narrative might look like, but it feels disheartening that the broad institutional response has been inevitable.”
CALI faculty reflection · Library Sciences

Şule Aksoy

Material Implications

“I do not want to lose sight of the damage that LLMs and big tech is doing at the human and planetary levels. So, while I want to teach students how to be thoughtful, ethical critics, I don’t know where I stand in terms of encouraging them to be ‘users’.”
CALI faculty reflection · English

“It’s important for students to reflect on the black box problem of AI and generative/predictive writing: what datasets were used to train these models? How do they reflect implicit and explicit biases, problems, discrimination, etc? If we can’t know these answers, how might we detect them from the texts these tools produce?”
CALI faculty reflection · Composition

Şule Aksoy

Implications for Science Education

Epistemic agency
Sensemaking
Interdisciplinary thinking
Pedagogy: SSI-based instruction

Şule Aksoy

Conclusion

“The danger I see with gen AI right now is just how far we are abstracted away from understanding the scale of planetary computation. Someone writing or coding or communicating in the so-called ‘cloud’ — as we are literally doing right now — does not need to envision the massive network of industrial processes, labor and power relationships, mineral extractions, deals between massive companies, etc. that continually weave the web that is the Internet. Gen AI only further scales this development, especially if students (or any of us) continue to lean on the tools. I worry that we are learning not to look at this massive infrastructure, at the materiality of the cloud.”
CALI faculty reflection · Computer Science

Frame #4 — Learning & Using STEM to Promote Justice (NASEM Equity in K-12 STEM Education, 2025)

Şule Aksoy

Conclusion

“this shift does make me feel that I can be more assertive by questioning my colleagues, professors, and future students when I see frivolous use of AI. Honest discussion about use of AI is especially important since my entire program and career path centers around reducing harm to marginalized communities. It is very difficult to be excited about new technology when I know my own students in an underserved urban area will be impacted most by its negative impacts.”
Preservice Elementary Teacher

Frame #5 — Envisioning Sustainable Futures Through STEM (NASEM Equity in K-12 STEM Education, 2025)

Envisioning sustainable futures — Erie Canal, water protectors, refusal imagery

NARST 2026

References & Further Reading

Bang, M., Warren, B., Rosebery, A. S., & Medin, D. (2012). Desettling Expectations in Science Education. Human Development, 55(5–6), 302–318.
Barad, K. (2007). Meeting the Universe Halfway: Quantum Physics and the Entanglement of Matter and Meaning. Duke University Press.
Bender, E. M., & Hanna, A. (n.d.). The AI Con: How to Fight Big Tech’s Hype and Create the Future We Want. Harper.
Brier, S., & Fabricant, M. (2016). Austerity Blues: Fighting for the Soul of Public Higher Education. Johns Hopkins University Press.
Cary, J., & Hull, G. (n.d.). Comprehensive AI Literacy: The Case for Centering Human Agency.
Cottom, T. M. (2025). Daring Ideas for the Future Series: Jason Reynolds and Tressie McMillan Cottom.
Crawford, K. (2021). Atlas of AI: Power, politics, and the planetary costs of artificial intelligence. Yale University Press.
Equity in K-12 STEM Education. (n.d.). National Academies of Sciences, Engineering, and Medicine.

Freire, P., & Macedo, D. (2000). Pedagogy of the Oppressed, 30th Anniversary Edition (M. B. Ramos, Trans.). Continuum.
Futch, V. A., & Fine, M. (2014). Mapping as a Method: History and Theoretical Commitments. Qualitative Research in Psychology, 11(1), 42–59.
Glitch as Infrastructural Monster. (n.d.). enculturation.
Gold, M. K., & Klein, L. (Eds.). (2023). Debates in the Digital Humanities 2023. University of Minnesota Press.
Guest, O., Suarez, M., Müller, B., van Meerkerk, E., Oude Groote Beverborg, A., de Haan, R., Reyes Elizondo, A., Blokpoel, M., Scharfenberg, N., Kleinherenbrink, A., Camerino, I., Woensdregt, M., Monett, D., Brown, J., Avraamidou, L., Alenda-Demoutiez, J., Hermans, F., & van Rooij, I. (2025). Against the Uncritical Adoption of “AI” Technologies in Academia. Zenodo.
Hao, K. (2025). Empire of AI: Dreams and Nightmares in Sam Altman’s OpenAI.
Haraway, D. (1988). Situated Knowledges: The Science Question in Feminism and the Privilege of Partial Perspective. Feminist Studies, 14(3), 575–599.

Harding, S. (1992). Rethinking Standpoint Epistemology: What Is “Strong Objectivity?” The Centennial Review, 36(3), 437–470.
Heeg, D., & Avraamidou, L. (2026). Demystifying AI: The urgency of a critical stance on the use of AI systems in education. PROSPECTS.
hooks, bell. (1994). Teaching to Transgress: Education as the Practice of Freedom. Routledge.
In Defense of the Arts, Humanities, and Social Sciences: A Call for a Transdisciplinary, Transnational, Ecological Approach to Science and Science Education. (2025). Journal of College Science Teaching, 54(4).
Lave, J., & Wenger, E. (1991). Situated Learning: Legitimate Peripheral Participation. Cambridge University Press.
Louridas, P. (1999). Design as bricolage: Anthropology meets design thinking. Design Studies, 20(6), 517–535.
Mader, A., & Dertien, E. (2016). Tinkering as Method in Academic Teaching. International Conference on Engineering and Product Design Education, Aalborg University, Denmark.
Morales-Doyle, D. (2019). There is no equity in a vacuum: On the importance of historical, political, and moral considerations in science education. Cultural Studies of Science Education, 14(2), 485–491.

Newfield, C. (2018). The Great Mistake. Johns Hopkins University Press.
Petrich, M., Wilkinson, K., & Bevan, B. (2013). It Looks Like Fun, But Are They Learning? In Design, Make, Play. Routledge.
Resnick, M., & Rosenbaum, E. (2013). Designing for Tinkerability. In Design, Make, Play. Routledge.
Ricaurte, P. (2019). Data Epistemologies, Coloniality of Power, and Resistance. Television & New Media, 20, 152747641983164.
Sano-Franchini, J. (2026). What’s “Critical” about “Critical AI”? A Recommitment to Humanistic Inquiry in the Ostensible March to Hyper-Automation. In C. D. M. Andrews, E. Kalodner-Martin, N. O’Connell, H. Wang, L. Wilkes, & C. Woods (Eds.), The Proceedings of the Computers and Writing Conference, 2025 (pp. 5–18). The WAC Clearinghouse.
Suarez, M., Müller, B. C. N., Guest, O., & van Rooij, I. (2025). Critical AI Literacy: Beyond hegemonic perspectives on sustainability.
Watters, A. (2021). Teaching Machines: The History of Personalized Learning. The MIT Press.

QR code linking to cuny.is/cali-narst — scan · cuny.is/cali-narst