Generative AI: Possibilities, Promises, Perils, Practices, and Policy

Download the report here: GenAI-5P

National Technology Leadership Summit
September 14-15, 2023
Washington DC

Report compiled by
Dr. Marie Heath
Assistant Professor of Educational Technology
School of Education
Loyola University Maryland

Dr. Punya Mishra
Associate Dean Scholarship & Innovation
Mary Lou Fulton Teachers College
Arizona State University

Introduction
The Generative AI, Possibilities, Promises, Perils, Practices, and Policy (genAI-P5) strand at
NTLS, led by Drs. Marie Heath and Punya Mishra, aimed to develop a set of questions for
teacher educators and teacher preparation programs to use for inquiring into policy, practice,
and research around generative AI. Participants explored the obvious and hidden impacts of
large language models (LLMs) on education and our individual and social lives. Participants
also applied technoskeptical (Krutka et al., 2022) and practice based questions (Mary Lou
Fulton Teachers College, 2023) to identify gaps in theory, positionality, and approaches to AI in
education. Finally, the strand participants developed a series of reflective questions to engage
with when considering the use of LLMs in education.

To generate the questions, participants reviewed diverse literature on the design and impacts of
LLMs (Bender et al., 2021) on education (Berkshire & Schneider, 2023; Heath & Krutka, 2023;
Mishra et al., 2023; Trust, 2023; Williamson, 2023), indigeneity (Hendrix, 2023; Marx, 2023), the
environment (Bender et al., 2021), and the social (Bender et al., 2021; Williamson, 2023). They
engaged in a technological audit (Krutka et al., 2022) of the technology and analyzed and
discussed the possibilities and perils. Next, the participants identified five important themes of
generative AI around which teacher educators can build further inquiry and reflection: truth and
verisimilitude; equity and justice; professional works, mindsets, tasks and skills; the broader
context of teacher preparation programs; teaching about Gen AI and its impacts on society.
Finally, participants brainstormed sub-topics and questions to ask about each theme.
Below are summaries of each theme and synthesis the participants’ brainstorms, including a set
of reflective questions for each theme.

Theme 1: Truth/Verisimilitude
Overview
The advent of generative AI, with its ability to craft realistic-looking synthetic media that can
easily be mistaken for reality, poses profound challenges for society at large, and thus becomes
relevant for educators. These technologies bring with them the potential for widespread
misinformation, as they can manipulate narratives, prioritize certain perspectives over others,
and even reshape our collective understanding of what we deem as truth, potentially altering our
very perceptions of reality. Added to this is the complex interplay of socio-economic factors, like
wealth and power which can influence the presentation and acceptance of algorithmically
produced ‘truths,” amplifying some narratives as “truth” while attempting to further marginalize
others and exacerbating existing schisms and inequities in our world. This evolution in AI
technology demands a heightened awareness and critical approach from educator preparation
programs to prepare the next generation of educators.

Reflective Questions

Does our curriculum/program help educators develop a critical understanding of the
nature of Generative AI and its ability to blur the lines between truth and falsehood,
subjective and objective truth?
a. How does our curriculum address the technological intricacies of Generative AI
that allow it to simulate reality?
b. Are educators introduced to discussions on the philosophical implications of
AI-generated truths versus human-derived truths?
Does our curriculum/program adequately prepare teachers to address the potential
spread of misinformation, and its impact on democratic society, given the ease of
creating realistic synthetic content?
a. How are teachers trained to identify and debunk AI-generated misinformation in
their classrooms?
b. Is there a component in the curriculum that delves into the broader societal
consequences of unchecked synthetic content on democratic processes?

Theme 2: Equity and Justice
Overview
While many technology companies and the powerful individuals who run them (including Musk,
Wozniak, and others) have speculated about future harms of generative AI on humans (Future
of Life Institute, 2023), algorithmic harms already exist in our present. Unlike the science fiction
dystopia presented by the open letter signed by Musk and other tech leaders, artificial
intelligence, or algorithmic models, currently cause material harm to people pushed to the
margins of society. Black feminist and queer scholars have called attention to the algorithmic
injustice embedded within the AI models and their damaging impact on marginalized people
(Benjamin, 2020; Costanza-Chock, 2020; Noble, 2018; O’Neil, 2017). For instance, algorithms
used in healthcare settings to determine medical interventions undercalculate the pain Black
women report, resulting in underdiagnosis and increased death compared to their white
counterparts (Benjamin, 2019). Despite a hope that AI would diminish or eliminate bias in
mortgage lending, AI reproduces housing and lending inequities, prompting lenders to reject a
higher percentage of loans and charge more interest to Black and Latinx applicants than their
white counterparts (Bartlett et al., 2022). Because of racism encoded in algorithmic learning,
Black people have been falsely arrested on the basis of poor facial recognition (e.g. Robert
Williams arrest by Detroit police; Kentayya, 2020; and Porcha Woodruff’s false arrest by Detroit
police; Cho, 2023). Trans people boarding planes are forced to walk through body scanning
systems which do not recognize their bodies, resulting in increased and invasive body searches
(Costanza-Chock, 2020). These are not potential harms of AI, they are existing harms that have
been occurring for years, despite the attention called to them by activists and scholars. Similar
biases have been seen in the use of ChatGPT in educational contexts as well (Warr, Oster, &
Isaac, 2023).

The rapid technological changes of generative AI, coupled with a hasty implementation in
education, may result in direct harms to already marginalized and minoritized students. How can
we work toward just uses of generative AI in education?

Reflective Questions:

How does generative AI currently and potentially intersect with systems of power in
education?
a. How are we preparing teachers to critically examine marginalized and minoritized
people’s lived experiences with generative AI?
b. How are we preparing teachers to identify systems of oppression which may be
amplified by using generative AI?
How does our TPP consider what facilitates and prevents access to generative AI in
educational spaces?
a. Are there tiered systems for access (free and paid)?
b. What other technology and resources are needed to access generative AI
models?
How do we evaluate whether generative AI is ethically designed for education use?
a. How is data collected and stored?
b. What is the aim of generative AI?
Does our curriculum/program emphasize the development of critical thinking skills to
interrogate whose perspective and narratives are being prioritized or marginalized by
AI-generated content?
c. How does our curriculum guide educators in recognizing and understanding
potential biases embedded within AI tools and outputs?
d. Are there discussions and exercises aimed at understanding the power dynamics
at play when algorithms decide which narratives to prioritize?

Theme 3: Professional Works, Mindsets, Tasks and Skills
Overview
Generative AI is a protean and multidimensional technology, with a wide range of capabilities to
produce complex, unique outputs across a range of domains (from programming to visual art,
from poetry to science and more). That said, it works well only when steered by a
knowledgeable human who brings their expertise both of domains and of working with AI to the
table. These capabilities can be an immense boon to educators, allowing them to develop
creative curricula and assessments and for their students to utilize them in creative ways to
support their own learning (Henriksen, Woo & Mishra, 2023). These tools offer quicker
curriculum adjustments and the development of innovative pedagogical approaches and
assessment techniques.

The importance of the human in learning cannot be overstated. Clarifying — for both themselves
and for the profession — what learning means and what an educator’s role is in the learning
process, will be crucial for articulating when and how to use generative AI in education.
Educators need to critically engage with AI-generated materials, discerning their relevance and
application (Close, Warr, & Mishra, 2023). Educators need to develop a creative mindset that
allows them to explore, play and understand the possibilities and challenges of bringing these
technologies to educational contexts (Warr, Mishra, Henriksen, & Woo, 2023). They also need
to be alert to the unintended consequence of AI in education, in particular a further
deprofessionalization of the profession. Technology companies and financially strapped districts
may point to the “efficiencies” of technology which they may argue — as they have with older
technologies — that it is an equal substitute for teaching.

Reflective Questions

Does our curriculum/program address how GenAI potentially reshapes the teaching
profession and educators’ roles?
a. Does our curriculum/program offer strategies for educators to maintain their
agency in the face of GenAI advancements?
b. Does our curriculum/program stress the importance of soft skills and human
values to maintain human-centric pedagogies in the face of AI integration?
Does our curriculum/program equip educators to effectively integrate and critically
evaluate GenAI-generated content?
a. Does our curriculum/program train teacher candidates to critically assess GenAI
outputs for specific disciplines and educational contexts?
b. Does our curriculum/program impart essential skills or knowledge for educators
to adapt and revise GenAI-generated content?
Does our curriculum/program provide opportunities for educators to learn how to best
work with GenAI in to develop and enact curricular goals?
a. Does our curriculum/program explore how rapid prototyping with GenAI might
lead to creative and innovative pedagogical strategies?
b. Does our curriculum/program guide educators developing new forms of
assessment that truly get at student learning and cannot be subverted by
generative AI?

Theme 4: Broader context of Teacher Preparation Programs
Overview
The integration of General Artificial Intelligence (GAI) into teacher preparation programs
presents a transformative shift, influencing not only admissions and evaluation processes but
also the transparency and support systems essential for pre-service teachers and instructors.
As GAI technologies evolve and potentially become as commonplace as smartphones within the
next five years, it’s crucial to anticipate and strategically plan for their implications in teacher
education. Key considerations include the impact of GAI on the admission of teacher recruits.
This encompasses how AI might alter existing barriers, potentially streamlining the process or
inadvertently creating new hurdles, particularly for diverse candidates. Furthermore, the role of
AI in the ongoing evaluation of teachers during their training period is a vital aspect, raising
questions about the fairness and inclusivity of such assessments.

Equity in AI-driven assessment systems is a paramount concern, particularly in addressing
challenges related to language diversity, accents, multilingualism, and disabilities. It’s essential
to consider whether AI can effectively identify relevant capacities and dispositions in teacher
candidates without reinforcing existing biases or inequities. Transparency in the deployment of
these AI systems is critical, ensuring that all participants understand how their performance is
being assessed and the basis of the feedback provided. Additionally, exploring key technological
points of entry that allow for the integration of AI into the education system will be crucial in
managing its impact on teacher training. This includes considering what data the AI requires
and ensuring that the machine learning algorithms and advisory paths do not perpetuate
structural inequities within teacher education and the broader landscape of higher education.

Reflective Questions

How do we create equity in AI driven assessment systems?
a. Will AI be used to admit students to the program? How might AI increase barriers
for admission and how might it reduce unnecessary barriers to admission?
b. Will pre-service teachers be given feedback by AI systems? How will we ensure
equity for all, including multilingual students, disabled students, and other
students who may potentially be marginalized by the use of AI?
Will we use AI to advise students throughout their programmatic experience?
a. What data will the AI need?
b. How do we ensure that the machine learning and advisory path devised by the AI
will not reproduce structural inequities within teacher education and higher
education?

Theme 5: Teaching About Generative AI and Its Impacts on Society
Overview
Not only will educators need to consider if and how they will incorporate generative AI in their
teaching, they will also need to prepare students to live in a world shaped by generative AI
(Richardson, Oster, Henriksen, & Mishra, 2023). As young citizens engage with technologies in
their daily lives, children deserve a curriculum that allows them to think about the impact of
technology on themselves and their world. Technologies themselves can and should be
contested, subject to reconstruction and democratic participation (Feenberg, 1991). Teachers
can help students transition perspectives from passive users to active citizens who make
informed decisions and take action for more just communities. The curriculum of generative AI
which teacher preparation programs might consider implementing includes teaching with, about,
and against technologies (Yadav & Lachey, 2022) and towards a civics of technology (Krutka &
Heath, 2022) which helps students examine the force technology exerts on society and the
ways that technologies extend biases of society.

Reflective Questions

Does your program equip teachers to address what K-12 students will need to be able to
know and do in order to live within a world with pervasive LLMs?
a. What do students need to know about the specific technological workings of
generativeAI in order to make informed decisions about its presence and use in
their lives?
b. What do students need to know about the ways that generativeAI intersects with
and amplifies societal biases?
c. What might constitute ethical uses of AI in students’ daily, educational, and social
lives?
Which disciplines and grades can include standards to build knowledge and skills about
the social and ethical impacts of generative AI?
a. How can each of the disciplines bring their disciplinary lenses to a holistic
understanding of AI in society?
b. What are age and developmentally appropriate ways to teach about, with, and
against AI?
How can teacher preparation programs build these standards into their programmatic
curricula?
a. What is an iterative process for incorporating this content into courses in teacher
preparation programs?

NTLS participants in this strand that contributed to the development of this report: Nadia Behizadeh, Chair, English Language Arts Teacher Educators (ELATE); Jongpil Cheon, Associate Chair, SITE Information Technology Council; Jonathan Cohen, President, SITE; Richard Culatta, CEO, ISTE; Shernette Dunn, Assistant Chair, SITE Information Technology Council; Alison Egan, Associate Chair, SITE Consultative Council; Sam Farley, Director of Sales, GoReact; Teresa Foulger, Associate Chair, SITE Teacher Education Council; Merideth Garcia, Co-Chair, Digital Literacies in Teacher Education (ELATE); Mark Hofer, Director, Studio for Teaching and Learning Innovation, Tristan Johnson, Editor, Educational Technology, Research & Development (ETR&D); Elizabeth Langran, Past-President, SITE; John Lee, Chair Emeritus, SITE Teacher Education Council; Jennifer Lesh, Past-President, CEC; Lin Lin Lipsmeyer, Editor, Educational Technology, Research & Development (ETR&D); Don McMahon, Past-President, Innovations in Special Education Technology (ISET), Council for Exceptional Children; Natalie Milman, Editor, CITE Journal, Current Practice; Erin Mote, Executive Director and Co-Founder, Innovate EDU; Tara Nattrass, Senior Education Strategist, Dell Technologies; Andrea Prejean, Director, Teacher Quality, National Education Association; Mary Rice, Editor, Journal of Online Learning Research; Denise Schmidt-Crawford, Editor, Journal of Digital Learning in Teacher Education (JDLTE), Past-President SITE; Mike Searson, Past-president, SITE; Melanie Shoffner, Editor, English Education (ELATE); David Slykhuis, Chair, National Technology Leadership Summit; Ji Soo Song, Digital Equity Advisor, US DoE, Office of Educational Technology; Joseph South, Chief Learning Officer, ISTE; Guy Trainin, AACTE Committee on Innovation and Technology, Lucas Vasconcelos, Emerging Leader, SITE; Lorrie, Webb, Co-Chair, AACTE Committee on Innovation and Technology.

References
Bartlett, R., Morse, A., Stanton, R., & Wallace, N. (2022). Consumer-lending
discrimination in the FinTech era. Journal of Financial Economics, 143(1), 30-56.

Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021, March). On the
dangers of stochastic parrots: Can language models be too big?🦜. In
Proceedings of the 2021 ACM conference on fairness, accountability, and
transparency (pp. 610-623).

Benjamin, R. (2019). Assessing risk, automating racism. Science, 366(6464), 421-422
Benjamin, R. (2020). Race after technology: Abolitionist tools for the new Jim code.
Wiley.

Berkshire, J. & Schneider, J. (Hosts). (2023, August 8). AI Is Going to Upend Public
Education. Or Maybe Not. With Larry Cuban . In Have
You Heard.

Costanza-Chock, S. (2020). Design justice: Community-led practices to build the worlds
we need. The MIT Press.

Cho, K. (2023, August 7). Woman sues Detroit after facial recognition mistakes her for
crime suspect. Washington Post.
https://www.washingtonpost.com/nation/2023/08/07/michigan-porcha-woodruff-ar
rest-facial-recognition/

Close, K., Warr, M., & Mishra, P. (2023). The Ethical Consequences, Contestations, and
Possibilities of Designs in Educational Systems. TechTrends.
https://doi.org/10.1007/s11528-023-00900-7

Feenberg, A. (1991). Critical theory of technology (Vol. 5). Oxford University Press.
Future of Life Institute. (2023, March 22). Pause giant AI experiments: An open letter.
https://futureoflife.org/open-letter/pause-giant-ai-experiments/

Heath, M.K. & Krutka, D.G. (2023, April 23). Collectively Asking Techno Skeptical
Questions About ChatGPT. Civics of Technology Blog.

Hendrix, J. (Host). (2023, January 29). An Indigenous Perspective on Generative AI with
Michael Running Wolf . In Tech Policy Press, The Sunday
Show.

Henriksen, D., Woo, L. & Mishra, P. (2023). Creative Uses of ChatGPT for Education: a
Conversation with Ethan Mollick. TechTrends. DOI:
https://doi.org/10.1007/s11528-023-00862-w

Kentayya, S. (Director). (2020). Coded Bias. Netflix.

Krutka, D. G., Heath, M. K., & Smits, R. M. (2022). Toward a civics of technology. Journal
of Technology and Teacher Education, 30(2), 229-237.

Mary Lou Fulton Teachers College. (2023, DATE). 15 questions every college professor
should be asking about ChatGPT and other generative AI: Reflective slidedeck.
https://docs.google.com/presentation/d/1Yb6m_xJnx7hgzRronZ_gjjspHGv16jTZp
YCqkNBSQPY/edit#slide=id.g27dd2c7880a_0_6022

Marx, P. (Host). (2023, July 20). Big Tech Won’t Revitalize Indigenous Languages with
Keoni Mahelona . In Tech Won’t Save Us.

Mishra, P, Warr, M, & Islam, R. (2023): TPACK in the age of ChatGPT and Generative
AI. Journal of Digital Learning in Teacher Education, DOI:
10.1080/21532974.2023.2247480

Noble, S. U. (2018). Algorithms of oppression: How search engines reinforce racism.
NYU Press.

O’Neil, C. (2017). Weapons of math destruction: How big data increases inequality and
threatens democracy. Crown.

Richardson, C., Oster, N., Henriksen, D., & Mishra, P. (2023). Artificial Intelligence,
Responsible Innovation, and the Future of Humanity with Andrew Maynard.
TechTrends https://doi.org/10.1007/s11528-023-00921-2

Trust, T. (2023, April). ChatGPT and Education [Google Slides].

Warr, M. and Oster, N., & Isaac, R. (2023). Implicit Bias in Large Language Models:
Experimental Proof and Implications for Education. Available at SSRN:
https://ssrn.com/abstract=4625078 or http://dx.doi.org/10.2139/ssrn.4625078

Warr, M., Mishra, P., Henriksen, D., & Woo, L. J. (2023b). A chat about GPT3 (and other
forms of alien intelligence) with Chris Dede. TechTrends. DOI:
https://doi.org/10.1007/s11528-023-00843-z

Williamson, B. (2023). The Social life of AI in Education. International Journal of Artificial
Intelligence in Education, 1-8.

Yadav, A., & Lachney, M. (2022). Teaching with, about, and through technology: Visions
for the future of teacher education. Journal of Technology and Teacher
Education, 30(2), 189-200.