Guidance on the Future of

Computer Science Education

in an Age of AI

teachai.org/cs
Note to the Reader

This is a limited print for the launch of Guidance on the Future of Computer
Science Education in an Age of AI. We hope that you will engage in the
development of future briefs such as:

How can AI be used to broaden participation in CS?

What core CS concepts and practices might be emphasized or deemphasized
when we teach with and about AI?
How can we better support students with disabilities to learn with and about AI?
What can CS education learn from how AI is used in software development?
What AI-related skills and content topics should a fundamental CS experience
include?
How can CS teachers use AI to promote creativity, starting in primary grades?
How do we teach about AI in contexts with low technical infrastructure?
Guidance on the Future of Computer Science Education in an Age of AI

Partners
Guidance on the Future of Computer Science Education in an Age of AI was developed by TeachAI and the
Computer Science Teachers Association (CSTA) in partnership with a community of thought leaders.

Lead Partners

In partnership with Karen Brennan, Maya Israel, Shuchi Grover and Matti Tedre

Partners

TeachAI Steering Committee

In collaboration with the World Economic Forum

Guidance on the Future of Computer Science Education in an Age of AI

Table of Contents

1 Executive Summary

5 Why is it Still Important to Learn to Program?

11 How Are Computer Science Educators Teaching With and About AI?

How Can Students Become Critical Consumers and Responsible

19 Creators of AI?

The Role of AI in Computer Science Education:

27
Results from a Teacher Survey

Table of Contents
Guidance on the Future of Computer Science Education in an Age of AI

Executive Summary

Recent innovations in artificial intelligence (AI) are provoking a wide range of questions about the
powerful technology’s role in the computer science (CS) classroom. Educators seeking to understand the
opportunities and risks of AI are asking:

Why is it still necessary to learn to program?

What do students need to learn about AI?
How do we leverage AI to broaden access and participation in CS?

While AI offers possibilities for supporting teachers’ work and enhancing student learning, there are
also concerns about disparate societal impacts. Teachers are concerned about student privacy, the
overreliance on AI tools, and students not developing foundational CS skills. These risks demand a
cautious approach by everyone in the education community and a commitment to ensuring that AI
benefits all students. Guidance on the Future of Computer Science Education in an Age of AI is meant
to inspire the community to reflect on what it means to teach with and about AI in CS education.

The Guidance

At a time when the entire education community is The guidance can be used to support a broad audience:
grappling with how to realize the benefits of AI while Teachers
mitigating the risks, the briefs that compose Guidance Administrators
on the Future of Computer Science Education in an Curriculum Providers
Age of AI serve as the beginning of a discussion rather Professional Development Providers
than definitive answers. The briefs offer preliminary Standards Writers
insights for responsibly and effectively integrating AI Edtech Developers
into primary and secondary CS education, address Researchers
common misconceptions, and provide a balanced
perspective on critical issues.

Each brief culminates in reflection questions, audience-

specific calls to action, and questions for further study.
The guidance illustrates an education community
grappling with what AI means for the future of their
discipline and serves as an example for other disciplines.

Executive Summary 1
Guidance on the Future of Computer Science Education in an Age of AI

The Process

Developed by TeachAI and the Computer Science development providers,

Teachers Association (CSTA), Guidance on the Future a lead group of researchers and organizations, and
of Computer Science Education in an Age of AI was a global working group of non-profit organizations,
informed by: teacher associations, and industry leaders.

72 TeachAI advisory committee and government The guidance also draws on a literature review of
agency participants, dozens of research articles and the contributions of
46 expert interviews, more than 360 educators worldwide who responded to
8 focus groups featuring over 100 teachers, our May 2024 CSTA/TeachAI survey. See the Research
researchers, and curriculum and professional Collection at teachai.org/cs.

Each Brief
The first set of briefs in the guidance addresses three of 3 How Can Students Become Critical Consumers
the most pressing questions in CS education today: and Responsible Creators of AI?
Many teachers have long advocated for teaching
1 Why is it Still Important to Learn to Program? about technology's ethical and societal impacts as
Despite their coding abilities, AI tools are far from a core component of CS education. Amid the rise
perfect, and learning to program lays the of Generative AI (Gen AI), CS teachers can help
foundation for students to develop the conceptual students develop the practices needed to become
understanding, agency, and dispositions needed critical consumers and responsible creators of AI.
to understand, use, and evaluate these tools and
their outputs. Future briefs in the series will consider questions such as:

2 How Are Computer Science Educators Teaching How can AI be used to broaden participation in CS?
With and About AI? What core CS concepts and practices might be
Learning CS is a natural opportunity for students emphasized or deemphasized when we teach with
to explore the benefits, limitations, and societal and about AI?
impacts of emerging technologies. Early initiatives How can we better support students with disabilities
suggest that teaching with and about AI in CS to learn with and about AI?
education has the potential to promote What can CS education learn from how AI is used in
computational thinking, demystify AI, and equip software development?
students to use AI tools responsibly. A mindful What AI-related skills and content topics should a
approach to implementing AI tools is critical to fundamental CS experience include?
realizing the potential benefits of AI while How can CS teachers use AI to promote creativity,
mitigating the risks. starting in primary grades?
How do we teach about AI in contexts with low
technical infrastructure?

Executive Summary 2
Guidance on the Future of Computer Science Education in an Age of AI

Additional AI in Education Resources

For those school systems ready to begin developing For those just starting to learn about AI, see
guidance on the responsible use of AI, the AI Guidance Foundational Policy Ideas for AI in Education for
for Schools Toolkit provides seven principles for resources designed for education leaders: What Is AI?,
realizing the benefits of AI while mitigating the risks. AI in Education and the Workforce, and Classroom
Perspectives on AI.

Acknowledgments
We thank those who contributed their time and We’d also like to recognize Allen Antoine, Owen
expertise to inform this guidance and look forward to Astrachan, Elaine Atherton, Carla Brodley, Center for
continuing to learn together what it means to teach with Computational Thinking and Design at Aarhus
and about AI in CS education. University, Edie Cheng, CS Access, Leigh Ann DeLyser,
Charlotte Dungan, Crystal Furman, Dan Garcia, Kinnis
This guidance was developed by TeachAI and the CSTA Gosha, Mark Guzdial, William G. Harris, Sallie Holloway,
in partnership with Karen Brennan (Harvard University), Lori Jacques, Antti Kiviniemi, John Kleeman, Jill
Quinn Burke (Digital Promise), Marie desJardins (AAAI), Kowalchuk, Michael Littman, Yolanda Lozano, Don
Bruce Fuda (Grok Academy), Christina Gardner- Miller, John Mitchell, Rusty Nye, Jamie Payton, Chris
McCune (AI4K12 and University of Florida), Shuchi Piech, Jennifer Rosato, Mehran Sahami, Vicky
Grover (Looking Glass Ventures and Edfinity), Maya Sedgwick, Ben Shapiro, Sepehr Vakil, Sara Vogel,
Israel (University of Florida), Irene Lee (Everyday AI), Hannah Walden, Shira Wein, Benji Xie, and Aman
Matti Tedre (University of Eastern Finland), and Jan Yadav for their input.
Vahrenhold (German Informatics Society).
TeachAI is led by a steering committee of Code.org,
We thank the TeachAI Future of CS Workgroup, which ETS, ISTE, Khan Academy, and the World Economic
includes Adobe, Amazon, Black in AI, College Board, Forum.
Data Science 4 Everyone, Digital Promise, Education
Above All, ETS, Texas Advanced Computing Center, Authors:
GitHub, Google, Grok Academy, Infosys Foundation Pat Yongpradit, Code.org
USA, Institute for Advancing Computing Education, Caron Carlson, Code.org
ISTE, Microsoft, National Council of Teachers of Veronica Ellis, Code.org
Mathematics, NCWIT, RobinCode, and participating Bryan Twarek, CSTA
U.S. and global government agencies. Jacob Koressel, CSTA

Suggested Citation
TeachAI & CSTA (2024). Guidance on the Future of Computer Science Education in an Age of AI. Retrieved from
teachai.org/cs. [date].

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

Executive Summary 3
Guidance on the Future of Computer Science Education in an Age of AI

Why is it Still Important to Learn to Program?

Despite the ability of Generative AI (Gen AI) tools to write code, learning to program remains crucial for
developing the conceptual understanding, agency, and dispositions necessary to use AI tools effectively,
evaluate and modify AI-generated code, and understand the broader context and implications of
programming. Programming produces more than code; it encourages creativity, critical thinking, and
collaboration and lays a foundation for learning the discipline of computer science (CS). AI might
augment learning but cannot replace the understanding and skills gained through learning to program.

Recent advances in Gen AI’s ability to write code have Code sense supports the ability to debug effectively,
led some industry leaders to claim that learning to optimize performance, and understand the broader
program is obsolete. Even as professionals use AI tools context of the code within a system.
to generate a growing proportion of code, learning to
program continues to be a context for developing the “In my vision, the child programs the computer
problem-solving and computational thinking skills to use and, in doing so, both acquires a sense of
these tools appropriately, evaluate and modify their mastery over a piece of the most modern and
output, and place the results in context (Salehi et al., powerful technology and establishes an
2020). Rather than diminishing the importance of intimate contact with some of the deepest ideas
learning to program, Gen AI tools highlight the need for from science, from mathematics, and from the
a foundational understanding of programming to art of intellectual model building.”
comprehend and appreciate these tools.
- Seymour Papert, Mindstorms: Children,
Computers, And Powerful Ideas, 1980, p.5
In the same way that students must learn “number
sense” – a conceptual understanding of numbers and
how they are related and connected – before they can For our purposes, “programming” is used
move on to algebra or calculus, they need to develop interchangeably with coding and includes aspects of
“code sense” to be successful in this new era of AI- software development. Programming is defined as the
assisted programming (Johnson, 2024). creative process of designing, writing, testing, and
maintaining code to perform specific functions and
solve problems.
Code Sense
In this brief, we dispel the myth that Gen AI coding tools
We define code sense as the conceptual make it unnecessary for students to learn to program.
understanding of a computer program's underlying
design, processes, and system relationships, as well
as the mental capacity to analyze, simulate, and
predict a program’s behavior.

Why is it Still Important to Learn to Program? 5

Guidance on the Future of Computer Science Education in an Age of AI

Learn to Program or Follow Commands: A Comic

AI is not magic. It needs human expertise and

guidance.
Myth 1: AI coding tools make programming knowledge understanding of algorithms and data structures. This
and skills unnecessary. knowledge is necessary to direct AI tools effectively.

To maximize the value and effectiveness of AI tools, For example, early research studying novices aged 10 to
developers need to know how to structure effective 17 years using code generators has shown that those
prompts, refinements, and requests (Kirova et al., with greater programming fluency benefit more from
2024). Learning to program, which is more than the tool. This suggests that AI tools are most effective in
learning the syntax of a programming language, gives augmenting, rather than replacing, programming
students opportunities to understand programming knowledge and skills (Kazemitabaar et al., 2023).
logic, practice solving problems, and develop an

Why is it Still Important to Learn to Program? 6

Guidance on the Future of Computer Science Education in an Age of AI

How Will Programming Evolve in

an Age of AI?

From assembly language to modern block-based

languages, programming has continuously evolved
to be more accessible by abstracting lower-level
programming details. Gen AI tools may continue “We need to help students learn to ask
this trend and allow programmers to focus more questions well. We need to teach them to have
on creativity, problem-solving, and addressing ownership of these tools.”
ethical concerns rather than the mechanics of
translating instructions into a programming - Christina Gardner-McCune,
language. Some programming tasks will be Associate Professor, University of Florida and
augmented by AI, and some will be replaced. Co-Lead of AI4K12

How should learning to program evolve to reflect

the changing nature of programming itself? What
insights can CS education gain from how software
engineers use Gen AI tools? Sign up for updates to
learn more about the future of CS education.

AI is not perfect. It needs human oversight.

Myth 2: Students do not need to learn to program “AI makes mistakes. Students need to
because AI can do it accurately and independently. understand the basics first, so they can critically
understand what they need from AI and also
Although Gen AI tools are improving, they are known to how to evaluate the results.”
suggest code that is inaccurate, contains security
- Neda Blackburn, STEM Director, Roland Park
vulnerabilities and biases, or is misaligned with a user’s
Country School
intent. A 2023 study of three popular code generators
found that they produced correct code between 31%
and 65% of the time (Yetiştiren et al., 2023). Another
Survey Insight
study showed that human and AI pairs critique code
more effectively than either can on their own. Learning CS teachers responding to a May 2024
to program lays the groundwork for developing the CSTA/TeachAI survey commonly noted that even
code sense and computational thinking skills needed to with AI's prevalence and coding abilities, learning
evaluate, debug, modify, and situate AI-generated to program is still essential for students to foster
code in the context of a larger program. the code sense and computational thinking skills
necessary to use and build future AI applications.

Why is it Still Important to Learn to Program? 7

Guidance on the Future of Computer Science Education in an Age of AI

AI will not replace the need for programmers.

It needs human creativity and domain expertise.

Myth 3: Students do not need to learn to program Human creativity and domain expertise have been and
because AI will replace all programming jobs. will be crucial in programming. Creativity drives
innovative problem-solving, intuitive user interface
The rapid uptake of Gen AI in the workplace is changing design, and feature development, while domain
programming practices without obviating the need for expertise ensures the solutions are accurate, relevant,
humans. Learning to program is shifting from focusing and address specific use cases. Together, they enable
on code generation to more code reading, evaluation, the creation of robust, user-friendly programs that
debugging, and refactoring (Kirova et al., 2024). meet both technical and contextual needs.

Focus on the Fundamentals “There is a longstanding need for people who

have software development skills. What you
Gen AI coding tools are rapidly improving to help need to know may shift – things that were
software engineers by eliminating repetitive tasks, obviously critically important or seen as
suggesting code, and even translating code from peripheral will need to change in their relative
one programming language to another (Stryker, importance.”
2024). While industry professionals will continue to
adapt to these innovations, students learning to - Ben Shapiro, Associate Professor, University
program need to build the capacity to understand of Washington
the fundamental, durable concepts behind these
evolving tools to use them appropriately.

AI can code, but there is more to programming than code.

Myth 4: The only purpose of learning to program is to

produce programs; now, AI can do that for us.

Programming goes beyond translating instructions into When students work together, they may find it easier
a programming language; it involves exploring needs, not only to absorb new concepts and vocabulary but
writing specifications, and testing, debugging, and also to express their curiosity and enthusiasm. Research
analyzing code. Learning to program offers students a shows pair programming can boost student retention in
platform for collaborative problem-solving, creative CS, particularly among some underrepresented groups
expression, and discovering joy in creating something
new (Kafai et al., 2015).

Why is it Still Important to Learn to Program? 8

Guidance on the Future of Computer Science Education in an Age of AI

(Campe et al., 2019). Having these opportunities while

Computer Science is More than
learning to program nurtures resilient, life-long learners
Programming
equipped to thrive in a rapidly changing world.

Learning to program is just one aspect of learning

Learning to program also has emotional value. A small-
CS. For example, the CSTA K-12 Standards also
scale study of primary school students showed that
include understanding how computers work, how
children felt significantly happier, more excited, and
messages are sent through the Internet, how data
more in control after learning to program (Tisza et al.,
is collected and analyzed, and how technology
2023). Results like these depend on many factors,
impacts society. Even if programming were
especially a high-quality teacher, and they highlight the
obsolete, a foundational CS experience would still
benefits learning to program can have on students’
be essential to:
emotional state, reminding us to consider how AI use
fueling innovation across multiple fields, such as
may affect students’ feelings about CS.
data science or computational biology,
addressing cultural and structural barriers to
pursuing post-secondary CS,
“Learning how to code isn't just about
expanding access to high-paying, highly
preparing young people for potential
satisfying jobs, and
participation in the technology workforce.
promoting greater diversity in a myriad of
Learning how to code can be an incredible
fields (Lewis, 2017).
opportunity for creative expression, enabling
learners to share their ideas and voice with the
world, bringing joy, satisfaction, and a sense of
purpose.”

- Karen Brennan, Professor, Harvard

University

Why is it Still Important to Learn to Program? 9

Guidance on the Future of Computer Science Education in an Age of AI

Three Ways to Take Action

Why is it Still Important to Learn to Program?

1. Given that AI can generate code, how can you balance the use of AI tools with the need
for students to develop fundamental coding skills and code sense?
Reflection 2. In what ways can you explicitly emphasize computational thinking skills, whether teaching
students programming, writing curriculum, or training teachers? What practical activities
or resources can help?

Administrators
1. Continue offering introductory CS courses that include programming.
2. Support CS teachers’ professional growth through communities such as a CSTA chapter.
3. Support schoolwide professional development and help teachers share lessons learned.
4. Develop guidelines based on resources like the TeachAI AI Guidance for Schools Toolkit.
Call to
Action Teachers
1. Be explicit about the fundamental concepts and skills you want students to learn (see
CSTA K-12 Standards), such as debugging, regardless of whether they use AI tools.
2. Explore using AI tools in the learning process before introducing them to students.
3. Discuss if, when, and how AI tools should be used in CS education. Develop principles to
guide AI use in programming assignments.

1. How does learning how to program with Gen AI tools compare to traditional instructional
methods? How might they help or hurt?
Further
2. How might coding assistant chatbots impact students with less prior CS experience in CS?
Study
3. What are the most common errors and biases found in AI-generated code in educational
settings? How can educators train students to identify, evaluate, and correct these issues?

TeachAI is led by Code.org, ETS, the International Society for Technology in Education,
Khan Academy, and the World Economic Forum.

TeachAI & CSTA (2024). Guidance on the Future of Computer Science Education in an Age of AI. Retrieved from teachai.org/cs.
Guidance on the Future of Computer Science Education in an Age of AI

How Are Computer Science Educators

Teaching With and About AI?

Computer science (CS) is a natural context for students to explore AI's benefits, limitations, and societal
impacts, and CS teachers overwhelmingly want professional development to facilitate these learning
opportunities effectively. Many teachers already teach with and about AI in a CS context to emphasize
computational thinking, demystify AI, and equip students to use AI tools responsibly.

As AI becomes pervasive in students’ lives in everything consider the appropriateness of AI tools for specific
from recommendation systems to voice assistants, CS activities even when the tools are permitted. There is a
teachers are incorporating lessons to demystify AI and growing body of unplugged resources to help students
encourage a critical mindset toward its benefits, learn about AI when policies or limited technological
limitations, and societal impacts. CS teachers are using infrastructure prevent AI use.
AI tools to help explain code, provide timely feedback to
students, and help them break through “coder’s block” In this brief, we examine how teachers teach with AI,
and syntax challenges so that they can focus on the such as using AI as a tool for student learning or teacher
creative aspects of programming. All teachers should support, and teach about AI, such as how it works and
follow existing policies, such as age restrictions based its ethical and societal implications.
on privacy or security concerns, and

Survey Highlights

Guidance on Computer Science Education in an Age of AI is informed by a survey of CS teachers (n = 364 teachers,
24% primary, 76% secondary, 12% international) administered by the CSTA and TeachAI in May 2024.

85% 48% 88%

of teachers think students in of teachers said they feel of teachers said they would benefit
introductory courses should learn equipped to teach about AI. from professional development to
about AI. learn how to use and teach about AI.

How Are Computer Science Educators Teaching With and About AI? 11
Guidance on the Future of Computer Science Education in an Age of AI

mathematics, language, biology, and social science

Teaching About AI (Casal-Otero et al., 2023).

“I do hope that educators realize the

CS teachers are incorporating AI into
importance of facing the reality of AI as well as
discussions, demonstrations, and unplugged
understanding that the goal of a paradigm shift
activities to demystify AI and make it more
like this should not be about policing students
approachable.
but about enabling students to think about using
[and] understanding this new technology.”
CS teachers are teaching students how AI works and
discussing AI’s ethical implications. Most teachers (79%) - Ben Smith, CS Teacher, Rowland Hall School
agree that curriculum and standards should be updated
to emphasize AI.

AI4K12
How much do students currently learn
Since 2018, the AI4K12 Initiative has been
about AI in your computer science curriculum?
developing national guidelines for teaching about
AI, organized around the 5 Big Ideas in AI.
0 hours

1-2 hours

3-5 hours

6-19 hours

20+ hours

0% 10% 20% 30% 40% 50%

Teachers indicated that foundational CS concepts like

algorithms help students develop the problem-solving
skills and awareness of societal impacts required to use
AI tools critically and to create with them responsibly.
Classroom activities about AI can examine topics such
as:

underlying programming and algorithmic concepts,

common uses, AI in Primary CS Education
limitations, and
ethical use and impact on society (Lee & Kwon, Vicky Sedgwick, a primary school educator, integrates
2024). machine learning into lessons that address the CSTA
K-12 Standards on data and analysis. Students sort,
What’s more, topics such as algorithms, classifiers, classify, and label data and use the data to train a
natural language processing, and AI’s ethical issues simple model. Students can then predict how the model
extend beyond the CS classroom and are relevant in
How Are Computer Science Educators Teaching With and About AI? 12
Guidance on the Future of Computer Science Education in an Age of AI

might classify new pieces of data and explore how Revised CSTA K-12 Standards, slated for release in
machines learn and begin to recognize bias in data. AI summer 2026, will include AI-related learning goals. In
concepts like natural language processing and image the meantime, the Reimagining CS Pathways project
recognition also fit into algorithms and programming has defined foundational CS content that includes
standards. Elementary and middle school students can learning progressions specific to AI.
explore how AI works with resources such as Code.org’s
AI for Oceans, Machine Learning For Kids, or MIT RAISE “[T]here is a tendency to make AI seem either
magical, sentient, infallible, or overly human . . .
Playground.
Since such (mis)representations are rife in
mainstream discourse, K-12 education needs to
There is also a growing body of unplugged activities
work extra hard to address this challenge
that do not require access to AI or a computer and are
through approaches to demystify AI and lift the
available for teaching AI concepts. For example, AI
hood on how it works .” (Grover, 2024)
technology playing cards teach about input, algorithm,
and output to demonstrate how developers’ choices - Shuchi Grover, Director, Looking Glass
affect results (Long et al., 2021). Ventures and Edfinity

Teaching with AI

CS teachers are integrating AI tools into only helped students develop an understanding of AI
instruction, assignments, and assessment. concepts but also made learning more fun. Although the
specific way that AI is used in the classroom is an
A systematic review of K-12 AI education from 2018 to essential factor, the review also suggested that using AI
2023 describes a variety of early initiatives to teach in the classroom can help students:
with AI by using machine learning tools and AI platforms
to help students understand basic concepts, such as develop foundational CS skills,
training data, testing models, and identifying bias (Lee reflect on critical and ethical considerations about
& Kwon, 2024). Hands-on experiences with AI tools not the role of AI and its use,
apply AI knowledge to real-life problems, and
improve motivation and interest in technology.

Thinking about ways AI tools could be used

in CS education…

52% said the 43% said the

potential potential
benefits benefits and
outweigh risks are equal
the risks

5% said the risks outweigh

the potential benefits

How Are Computer Science Educators Teaching With and About AI? 13
Guidance on the Future of Computer Science Education in an Age of AI

AI Feedback Learning from Math Education

Maria Camarena, a secondary CS teacher in California, Lori Jacques, a former math teacher and current CS
uses ChatGPT to improve students’ programming skills. professor, describes three approaches based on math
Students input lines of code they have written and a pedagogy to help students look at problems differently
description of the code and receive programming and and improve their programming skills and conceptual
writing feedback. This approach strengthens understanding (Jacques, 2023):
programming skills and supports multilingual students
with their English language development. Create Multiple Representations: Students can use
Generative AI (Gen AI) to generate code, explain
Early research also identifies opportunities to use Large code in plain language, illustrate how the output
Language Models (LLMs) to improve the CS learning changes as the code executes using a flow chart,
experience and foster critical-thinking skills, including: and compare the different representations and
their pros and cons.
generating code as a starting point for a solution,
teaching students how to describe solutions in Explore Different Approaches: Students can use AI
natural language, and code generators to create different programs to
guiding problem-solving strategies (Prather et al., address the same problem and then compare them
2023). to each other and human-generated solutions,
including their own.
LLMs have also been shown to help provide a vast
selection of solutions of varying quality so students can Explain Code Created by AI: Students can analyze,
practice critical analysis by examining multiple answers describe, and explain the functionality of AI-
to a problem (Denny et al., 2023). generated “worked examples,” which may or may
not be correct.
In one early effort to revise a post-secondary
introductory CS curriculum to use LLMs, problem
decomposition and explaining/testing/debugging code
were emphasized over syntax and writing code
(Vadaparty et al., 2024). The scope of students’
projects transcended what was typically seen in a CS1
course, and their exam results reflected results similar
to previous CS1 classes.

Pairing with AI

Michael Phelan, a CS teacher in North Carolina, allows

students in his Python 2 class to use ChatGPT or Copilot
to give them an experience similar to pair programming
while pushing them to improve their own thinking.
However, it is vital to Phelan that students have a
comprehensive understanding of what they are doing
rather than relying on AI to do the work.

How Are Computer Science Educators Teaching With and About AI? 14
Guidance on the Future of Computer Science Education in an Age of AI

Concerns About AI’s Risks “Students have a tendency to 'trust' the AI and
blindly copy and paste the code but lack the
CS teachers responding to the CSTA/TeachAI survey skills to 'talk to' the code with AI and truly
frequently cited these three concerns about the understand what is going on. This causes them
potential risks of using AI in the classroom: frustration when later attempts don't work, and
they lack the vocabulary or experience needed
to properly explain what problems or tasks
AI will do the thinking for students.
need to be done to their AI assistant.”
Students will become overreliant on AI tools.
Students will lose essential skills. - Graham Nolan, CS Teacher, Hong Kong
International School
These concerns can be mitigated by teaching
foundational programming skills and critical
engagement with AI tools. This approach ensures that
students use AI as a supportive tool, maintaining their For more on the ongoing need for students to learn CS
problem-solving abilities and understanding the fundamentals, see Why is it Still Important to Learn to
underlying logic. Program?

Teacher Support

CS teachers are using AI to support and

streamline their own work.

Teachers who participated in the CSTA/TeachAI survey

reported using AI to facilitate their work in various
ways, including tailoring and refining lesson plans;
providing differentiation for students who need either
more challenge or more support; and organizing ideas,
breaking them down, and re-explaining them.

These uses mirror common opportunities for Gen AI to

“While AI can automate certain tasks, such as
streamline teacher workloads. In one early study,
grading or feedback provision, overreliance on
researchers prompted the code generator Codex with
AI tools may diminish educators' role in the
source code to generate new programming exercises.
teaching process. It's essential to strike a
They found that a large majority were completely novel
balance between leveraging AI for efficiency
and aligned with the topic and theme at hand (Denny et
and maintaining the human element of
al., 2023). In addition to saving instructors’ time, the teaching, including personalized interaction and
variety of solutions that code generators can create mentorship.”
from one input can offer students a view into different
approaches to tackling the same problem. - Fred Sagwe, CS Teacher, Shimo La Tewa
School and Robotics Society of Kenya

How Are Computer Science Educators Teaching With and About AI? 15
Guidance on the Future of Computer Science Education in an Age of AI

Inaccuracy and Overreliance

Global Perspectives

Despite the potential benefits of using AI tools to As of July 2024, 41 state education agencies in the
enhance learning and provide instructor support, U.S. and over 25 government agencies globally
common concerns involve Gen AI's tendency to produce participate in TeachAI. Here are some global
inaccurate information and students’ ability to make perspectives on the future of CS education:
sense of the output. Both students and teachers also
express concerns that students may not learn their
“When we teach students how to understand
material when they turn to Gen AI for help (Zastudil et and responsibly leverage AI, we equip the next
al., 2023). generation of technology users and creators
with skills and confidence to innovate.”
Teachers have mixed opinions about whether students
in introductory CS classes should learn to program with - Oksana Pasichnyk, CS Teacher, Lyceum
AI. The benefits and risks of learning to program with AI Sykhivsky, Ukraine
tools are still being discovered and will depend on the
classroom context and how the tools are used. “AI in coding classrooms faces challenges in
implementation due to varying teacher and
student understanding and skill levels. Effective
Students in introductory CS classes should AI integration requires teacher and student
learn to program with AI. training, well-designed learning experiences,
ethical guidelines, and ongoing support. Despite
55% agree 21% disagree these challenges, AI offers immense potential to
transform coding education in Thailand.”

- Cheeraporn Sangkawetai, Institute for the

Promotion of Teaching Science and Technology,
Thailand
24% unsure

How Are Computer Science Educators Teaching With and About AI? 16
Guidance on the Future of Computer Science Education in an Age of AI

Three Ways to Take Action

How Are Computer Science Educators Teaching With and About AI?

1. In what situations might teaching with and about AI be useful? In what situations might it
be inappropriate? Re
Reflection 2. Where can teaching about AI be integrated into the existing curriculum? Q
3. What changes can be anticipated in students’ engagement, motivation, or understanding
of programming concepts after introducing AI tools?

Teachers
1. Participate in ongoing professional development (PD) to deepen knowledge and skills in
teaching with and about AI. Find PD from CSTA, CSTA chapters, and curated programs.
2. Follow local guidance on the use of AI tools in classrooms. When local policies do not exist,
develop guidelines for using AI tools on classroom assignments. Prioritize responsible
practices that encourage students to determine if, how, when, and why they would use AI. Ca
3. When using AI in your work, constantly evaluate the output to see how it fits your Ac
classroom context, objectives, and teaching style.
Call to
Action
Curriculum and Professional Development Providers
1. Include AI as a topic and tool and explore uses of AI that will enhance activities, examples,
lessons, and projects. Proceed cautiously, with an eye towards safe and responsible use.
2. Provide professional development enabling teachers to experiment, share use cases and
insights, and field questions and concerns in their schools.
3. Ground questions and projects for students in real-world examples that also illustrate the
benefits and risks of AI. Fu
St
1. How should CS curriculum and standards be updated to include AI concepts and tools?
2. How can school systems address barriers to access, support effective pedagogical design,
Further and provide powerful learning opportunities to provide equitable access to AI education?
Study See Classroom Perspectives on AI (TeachAI, 2024) for more information on these issues.
3. How do the interactions between teachers, students, and AI tools influence the teaching
and learning process and outcomes?

TeachAI is led by Code.org, ETS, the International Society for Technology in Education,
Khan Academy, and the World Economic Forum.

TeachAI & CSTA (2024). Guidance on the Future of Computer Science Education in an Age of AI. Retrieved from teachai.org/cs.
Guidance on the Future of Computer Science Education in an Age of AI

How Can Students Become Critical

Consumers and Responsible Creators of AI?

Teaching students to become critical consumers and responsible creators of AI involves integrating
ethical and societal considerations into computer science (CS) education. CS teachers are incorporating
discussions about transparency, accountability, information accuracy, privacy, fairness, and ethical
design into their curricula. These lessons empower students to evaluate AI outputs critically, understand
the implications of AI on society, and design technologies that prioritize equity and responsibility. More
than any other subject, CS provides an opportunity to explore AI’s societal and ethical impacts by
learning how AI models are trained, understanding how they work, and developing their own models.

From producing deepfakes to amplifying gender, racial, must also learn to ask whether and why they should use
and cultural biases, AI’s potential to harm individuals AI in the first place.
and communities is in the spotlight, creating increasing
demand for incorporating societal and ethical “When it comes to AI education, we do not have
considerations into CS education. the luxury of burying our heads in the sand. CS
teachers have the opportunity and
An essential component of teaching students about AI is responsibility to lead students in understanding
empowering them to understand how AI models are the societal and ethical implications of AI: the
built and to think critically about the benefits, harms, good and the bad, the benefits and harms, the
possible safeguards, and the people and communities possibilities and realities.”
affected by this powerful technology. As more CS
- Charity Freeman, CSTA Board Chair
education involves consuming and creating AI, students

Critical Consumers, Responsible Creators

Hmmm, am I Let me just make my

comfortable own photo filter app!
sharing my
images with
this app?

How Can Students Become Critical Consumers and Responsible Creators of AI? 19
Guidance on the Future of Computer Science Education in an Age of AI

According to Reimagining CS Pathways: High School

and Beyond, the societal impacts of computing and How often do you supplement your CS
related ethical implications are the highest-rated topics curriculum to include discussions of ethical
of importance across the CS education community. questions related to AI, such as algorithmic
bias, digital citizenship, and privacy?
Additionally, 66% of CS teachers responding to a
CSTA/TeachAI Survey (May 2024) said that Rarely or Never
conversations about AI have helped students
understand the societal and ethical issues surrounding Monthly
AI. Ethical concerns about AI also emerged as one of the
most requested professional learning topics.
Weekly (1-3 times)

In this brief, we highlight a few of AI’s ethical and

societal considerations and offer some practices for Every day

developing critical AI consumers and responsible AI

0% 10% 20% 30% 40%
creators.

Transparency and Accountability

AI systems are widely perceived as opaque, When asked about the potential risks of including AI in
containing unknowable data sources and the classroom, CS teachers responding to the
inscrutable decision-making processes. CSTA/TeachAI Survey emphasized that students must
be made aware that AI makes mistakes and can have
When equipped with explanations about how AI systems disparate societal impacts, often perpetuating harm on
work, teachers are less reluctant to use them in their marginalized communities. These same concerns were
classrooms and less concerned about the technology’s commonly cited as reasons for students to deepen their
impact on personal autonomy. Early research suggests core CS knowledge.
that teachers are more likely to trust AI when given the
ability to review and override AI recommendations
"Students are more empowered to engage in
(Nazaretsky et al., 2022).
popular AI discussions when they understand
the technical side of the ethical and societal
impacts of AI."

- Sepehr Vakil, Associate Professor,

Northwestern University
Classroom Practice: Create model cards that
explain the intended use, training dataset,
accuracy, and limitations of a classification or
prediction model.

How Can Students Become Critical Consumers and Responsible Creators of AI? 20
Guidance on the Future of Computer Science Education in an Age of AI

Information Accuracy and Integrity

AI makes it easier for unethical parties to specific artists to deepfakes in the democratic process,
produce misinformation and disinformation by which raise significant ethical and legal concerns about
mimicking voices, images, and videos. originality, consent, and the potential for
misinformation.
During the 2023-24 academic year, parents and
students in California, New Jersey, and Washington
sounded the alarm over AI: Classmates used AI-based
image generators to create explicit “deepfakes” from
pictures online and circulated the images on social Classroom Practice: Discuss the importance of
media. Districts responded by updating their social verifying information sources and explore ways to
media guidelines and educating students about the validate content. Explore instances of when AI
appropriate and acceptable use of Generative AI (Gen output adds value to individuals and society and
AI) tools while urging others to learn more about when it does not. Discuss the importance of citing
misinformation. There is a broad range of content sources when using AI.
authenticity issues, from creative works that imitate

"Students can potentially misuse AI for "The hard problems of AI and ML education
themselves and others, but teaching them and come down to civic values, philosophy, ethics,
being open about the benefits and harmful fairness, bias, transparency, explainability. The
effects and potential upsides and downsides far social studies teachers - we can learn from
outweigh the potential risks." them."

- Art Lopez, Project Manager, Coding Our - Ben Shapiro, Associate Professor, University
Future, University of California, San Diego of Washington

Privacy Rights

Data collected by AI-powered systems and surveillance, behavioral threat assessments, predictive
educational tools raise concerns about students’ policing, and other surveillance tools.
privacy rights.
As students turn over growing volumes of personal
In March 2024, a coalition of 41 educational and civil data, including data collected by educational tools, they
rights organizations wrote to the U.S. Department of must understand the potential for combining it to
Education, asking to end the funding of AI-powered divulge sensitive information. Moreover, tracking and
security systems in K-12 schools. They were concerned monitoring components in AI systems that predict
about AI’s potential to violate students’ civil rights. student performance, weaknesses, and strengths can
Especially alarming, the group said, was the increasing inhibit their participation at school (Akgun & Greenhow,
use of facial recognition technologies, social media 2021).

How Can Students Become Critical Consumers and Responsible Creators of AI? 21
Guidance on the Future of Computer Science Education in an Age of AI

"Anyone developing technologies has a

responsibility to keep ethics top of mind and
Classroom Practice: Discuss how, when, and why
create things that will not cause harm. There
students’ personal data are collected, analyzed,
should be a pledge or agreement that these
interpreted, and used. Consider how decision-
persons go through so that 'doing no harm' is at
makers might use this data to their advantage —
the basis of their profession. . . This must also
even if it is something students are not okay with.
be followed with action."
Discuss students’ rights and how they can opt out of
data collection. - DaQuan Bashir, Professional Learning
Manager, CSTA

Fairness and Justice

AI systems’ disparate impacts on different receive higher recidivism scores. A related issue raised
demographic groups raise alarms about fairness by the COMPAS example is transparency, as the
and justice in our digitally driven society. designers of COMPAS refused to reveal details of the
software and its algorithms.
AI systems can reflect and amplify the implicit biases of
their creators. Understanding AI biases involves
understanding how people create technologies like
Large Language Models (LLMs) and how biases can
creep into all stages of development, such as selecting Classroom Practice: Discuss the pros and cons of
training data and selecting which features in the using AI for different projects and consider when
training data to prioritize (Broll & Grover, 2023). another solution might be a better option. Discuss
Examples of discrimination against women, people who how AI tools might be used to detect and measure
are economically disadvantaged, or people of color as a bias in datasets.
result of algorithmic bias have shown up in hiring, health
care, insurance, and mortgage lending decisions, in
addition to criminal justice assessments. “[A]ll young people need to understand whether
the AI with which they knowingly or unknowingly
The non-profit journalism group ProPublica in 2016 was engage has treated them fairly.”
the first to uncover racial bias in COMPAS (Correctional (Holmes et al., 2022, p.31)
Offender Management Profiling for Alternative
Sanctions), an AI-powered system used to predict a
criminal defendant’s recidivism risk. COMPAS falsely
identified Black defendants as likely future criminals
nearly twice as often as white defendants. The AI model
was trained on historical crime data, and, as a result,
individuals who were part of a demographic historically
targeted by law enforcement were likely to

How Can Students Become Critical Consumers and Responsible Creators of AI? 22
Guidance on the Future of Computer Science Education in an Age of AI

Responsible AI & Tech Justice

The Kapor Foundation’s Responsible AI & Tech

Justice: A Guide for K-12 Education helps
educators develop learning experiences that
prioritize critically evaluating technologies and
creating more equitable solutions. One
recommendation is to “examine the AI technology
creation ecosystem from who designs and
develops products and how they are developed, to
who invests in their creation and benefits from
their adoption.”

Ethical Design

Transparency, accountability, equity, and

privacy should be principles in technology
development.
Classroom Practice: Early in the design process,
Ethical design recognizes that all technologies are have students explore different user personas:
embedded in social systems and reflect designers' representative examples of potential users with
beliefs. Ethical design considers societal values, includes different backgrounds, desires, ability levels, and
potential users throughout the design process, perspectives. For more information, see Practice 1:
especially those most affected by the design, and Fostering an Inclusive Computing Culture in the K-12
follows principles of transparency, explainability, CS Framework.
accountability, equity, and privacy.

In a study of recommendation algorithms and ethical

AI4K12: Societal Impact
design, students as young as 5th graders were shown to
be capable of recognizing design agendas and
AI is a disruptive technology that will not only
identifying influences on those agendas (DiPaola et al.,
impact the economy and employment but also
2020). Students were asked to identify relevant parties
shape social and cultural norms (AI4K12
associated with YouTube and their potential values,
Framework’s Big Idea 5).
encouraging questions about the data collected from
the platform, how it influences what they watch, and
how it affects their beliefs. Students then had an
opportunity to develop their own design goals and
agendas using ethical design principles.

How Can Students Become Critical Consumers and Responsible Creators of AI? 23
Guidance on the Future of Computer Science Education in an Age of AI

Ethical Concerns and Practices

Topic Ethical Concern Classroom Practice

Transparency and AI systems are widely perceived as Create model cards that explain the intended
Accountability opaque, containing unknowable use, training dataset, accuracy, and
data sources and inscrutable limitations of a classification or prediction
decision-making processes. model.

Information Accuracy AI makes it easier for unethical Discuss the importance of verifying
and Integrity parties to produce misinformation information sources and explore ways to
and disinformation by mimicking validate content.
voices, images, and videos.

Privacy Rights Data collected by AI-powered Discuss how, when, and why students’
systems and educational tools personal data is collected, analyzed,
highlight student privacy issues. interpreted, and used.

Fairness and Justice AI systems’ disparate impacts on List the pros and cons of using AI for different
different demographic groups projects and consider when other solutions
raise concerns about fairness and might be a better option.
justice.

Ethical Design Transparency, accountability, Address the needs of diverse users and
fairness, and privacy are crucial in consider when the risks of using AI might
technology development. outweigh the benefits for an individual or
community.

How Can Students Become Critical Consumers and Responsible Creators of AI? 24
Guidance on the Future of Computer Science Education in an Age of AI

Three Ways to Take Action

How Can Students Become Critical Consumers and Responsible Creators of AI?

1. How can you incorporate ethical and societal issues related to AI into CS content? How
can you address the potential challenges when engaging in sensitive topics? What can be
learned from other subjects?
Reflection 2. How can you model the ethical and responsible use of AI tools? What steps can students
take to understand AI tools' limitations and biases and evaluate their outputs critically?
3. What data do you and students share on online platforms and with AI systems, both
knowingly and unknowingly? What risks are associated with sharing personal data?

Standards Writers
1. Ensure that ethical considerations and societal impacts of AI and technology are
embedded in primary and secondary education standards.
2. Connect the societal and ethical impacts of AI to how AI works and how people develop AI
technologies.
3. Develop standards that integrate CS education with other disciplines, such as science,
social studies, philosophy, and ethics, to provide a holistic view of AI's societal impacts.
Call to
Action
Curriculum and Professional Development Providers
1. Include deliberate conversations about bias, ethics, and AI's impact throughout the
curriculum. Provide students with reflection questions at the start and end of each project.
2. Encourage perspective-taking by asking about user stories and how tools might work.
Frame the potential for positive, negative, and unintended consequences.
3. Create space for honest conversations and thoughtful questions. Include questioning
protocols to foster deeper thinking about how what people design will affect others.

1. How does incorporating ethical considerations, such as bias and fairness, into CS content
affect students’ understanding and critical analysis of AI technologies?
Further 2. What are effective methods for teaching transparency and accountability? How do these
Study methods influence students’ trust and willingness to engage with AI technologies?
3. How can educational interventions be designed to effectively reduce biases and promote
fairness in students’ AI projects and understanding?

TeachAI is led by Code.org, ETS, the International Society for Technology in Education,
Khan Academy, and the World Economic Forum.

TeachAI & CSTA (2024). Guidance on the Future of Computer Science Education in an Age of AI. Retrieved from teachai.org/cs.
Guidance on the Future of Computer Science Education in an Age of AI

The Role of AI in Computer Science

Education
Results from a Teacher Survey

Key Takeaways

1 CS teachers want to teach with and about AI. 4 CS teachers stress the importance of
teaching programming skills alongside AI
Most teachers (85%) believe using and learning education but disagree on when students
about AI should be included in foundational CS should start using AI tools.
experiences, and (80%) agree that CS standards
should include AI. Additionally, 87% say students Teachers believe programming is essential for
should learn about careers impacted by AI. understanding, debugging, and effectively
collaborating with AI and crucial for developing
computational thinking, problem-solving, and
2 CS teachers want support in teaching with critical thinking skills.
and about AI.
While 55% of CS teachers believe students in
A notable 88% of teachers expressed the need for introductory CS classes should learn to program
professional learning and additional resources to with AI, 24% are unsure, and 21% disagree.
effectively teach with and about AI, highlighting
the necessity for updated professional 5 CS teachers cite bias, misinformation,
development programs.
overreliance, and the ethical implications of
Gen AI technologies as common concerns.
3 CS teachers are mostly optimistic or neutral
about the potential benefits and drawbacks Across the survey, teachers identified the need to
of AI in CS education. talk with students about AI's ethical and societal
impacts. They expressed the desire for guidance
While 52% believe the benefits of AI outweigh the
on discussing these issues and how AI tools will
risks, 43% see an equal mix, and only 5% said the
specifically impact the classroom.
risks outweigh the benefits. Key concerns include
overreliance on AI, the potential for plagiarism,
and ethical issues such as bias in AI systems.

The Role of AI in Computer Science Education: Results from a Teacher Survey 27

Guidance on the Future of Computer Science Education in an Age of AI

CSTA and TeachAI conducted a survey in May 2024 about computer science teachers’ usage of, beliefs about, and
attitudes toward AI.

364 responses from current CS teachers 12% outside of the US

24% primary school, 76% secondary school 44% urban, 43% suburban, 13% rural

CSTA will gather additional data on these questions in a teacher landscape survey in the fall of 2024.

AI Curriculum Content

Why teach with and about AI? What should we teach about AI?

Teachers emphasized that understanding AI is essential There is a strong consensus among CS teachers about
for students' future personal and professional lives in the importance of AI education:
the digital age. 76% agree that AI can be used to 87% agree that students should learn about careers
promote creativity in the CS classroom. impacted by AI.
79% believe that CS standards should be updated to
Teachers were mostly optimistic or balanced about the include discussions about AI.
benefits and drawbacks of using AI in the classroom. 85% think students in introductory courses should
learn about AI.

How should CS curriculum evolve?

Thinking about ways AI tools could be used
in CS education… Teachers indicated that students should continue
learning fundamental programming skills, including
52% said the 43% said the topics like variables, loops, conditionals and topics that
potential potential directly apply to AI, like data structures and algorithms.
benefits benefits and
outweigh risks are equal
Many strongly advocated for students to learn about
the risks the societal impacts of AI in their foundational CS
courses. Many also responded that they were unsure
which CS topics should be emphasized or changed.
5% said the risks outweigh
the potential benefits “There is no getting around the need for
humans to master order of execution, variables,
logic, conditionals, looping, functions, modular
design, and so on.”

- Secondary CS Teacher, Maine, USA

The Role of AI in Computer Science Education: Results from a Teacher Survey 28

Guidance on the Future of Computer Science Education in an Age of AI

“I don't think any topics should be removed. But “I think topics should include more about the
I think that real world applications and the mechanics of computer science, the changing
future need to be emphasized more… I believe job landscape, daily life intersection with
that AI needs to be included into the standards computer science, and AI fundamentals for
as soon as possible.” students.”

- Secondary CS Teacher, Pennsylvania, USA - Primary and Secondary CS Teacher, Ohio,

USA
“Collaboration should always be emphasized.
Struggling with a problem to find a solution
should be emphasized. Understanding that
there are multiple ways to solve a problem
should be emphasized.”

- Secondary CS Teacher, Maryland, USA

Teaching Methods and Support

How often are teachers teaching about AI?

CS teachers are teaching students how AI works and

discussing AI’s ethical implications with students.

How often do you discuss ethical

How much do students currently learn questions related to AI, such as algorithmic
about AI in your computer science curriculum? bias, digital citizenship, and privacy?

0 hours Rarely or Never

1-2 hours
Monthly

3-5 hours

Weekly (1-3 times)

6-19 hours

20+ hours Every day

0% 10% 20% 30% 40% 50% 0% 10% 20% 30% 40%

The Role of AI in Computer Science Education: Results from a Teacher Survey 29

Guidance on the Future of Computer Science Education in an Age of AI

When should AI be used in the classroom?

There is a division among teachers on when to introduce

AI tools. “I want to be able to spend my time gathering
new resources to use in my classroom and less
time grading papers and creating worksheets
Students in introductory CS classes should and exams.”
learn to program with AI.
- Primary CS Teacher, New Jersey, USA
55% agree 21% disagree

“For professional development tailored to

computer science teachers, focusing on AI
topics like machine learning fundamentals,
ethical considerations in AI development,
24% unsure integrating AI concepts into curriculum, and
practical coding applications would be crucial.”

- Secondary CS Teacher, Bauchi, Nigeria

Interestingly, 54% of secondary school teachers agreed
that students should learn to use AI in introductory
programming courses, versus 62% of primary school “I would love to see more PD for elementary
teachers. level CS teachers with ways to ready these
young learners without necessarily using Gen
What AI topics should be included in CS AI. Also, teachers need PD more readily
available for themselves. Our district thus far
teacher professional development?
has provided none, so unless a teacher is highly
motivated to seek understanding themselves,
Teachers want to integrate balanced discussions about they are not getting it.”
the benefits and risks of AI with hands-on projects, but
they require additional support to teach with and about - Primary and Secondary CS Teacher, Arizona,
AI effectively. USA

48% feel currently equipped to teach AI.

88% need more resources for professional learning.
4% disagree that they need additional support.

The most commonly requested topics for professional

development centered on how AI works, how to use AI,
and the ethical impacts of AI tools. While these topics
are not exclusive to the CS classroom, teachers
identified their relevance to CS learning and their
importance to developing digitally literate students.

The Role of AI in Computer Science Education: Results from a Teacher Survey 30

Guidance on the Future of Computer Science Education in an Age of AI

Student Learning and Development

Why should students still learn to program? “AI is no substitute for human ingenuity.
Students still need to learn problem-solving,
Teachers most commonly responded that given Gen AI logic, teamwork, communication, etc., which is
coding tools, students should still learn to program to be the heart of computer science.”
able to develop new Gen AI applications, understand - Secondary CS Teacher, Massachusetts, USA
how AI works, and ground their work in fundamental
computer science knowledge.
“Coding teaches students to think
computationally, problem-solve, think through
“Learning computer science is still crucial for
steps, cause and effect, debugging (finding and
students, even with the rise of AI.
fixing their mistakes), and critical thinking skills.
Understanding the principles behind coding and
These skills are not caught; they are taught and
computational thinking not only enables
learned.”
students to create AI but also helps them grasp
its limitations, ethical considerations, and - Primary CS Teacher, New Jersey, USA
potential societal impacts.”

- Secondary CS Teacher, New York, USA What are the potential benefits of using AI in
the CS classroom?
Teachers emphasized that developing a robust “code
sense” through learning to program is just as crucial
CS teachers most commonly identified the potential
with AI as without. Many teachers noted that AI does
benefits of using AI in their classrooms as increasing
not generate perfect code, and students must have a
learning opportunities, automating tasks, and preparing
solid foundation in programming to analyze, debug, and
students for their futures in the digital world.
use what AI creates.

“As with many parts of education, students They most commonly identified AI as helpful in
need to learn the mindset and logic of enhancing students' learning opportunities by
programming. This allows them to fix mistakes strengthening engagement, broadening access, and
or optimize poorly written code, whether that is shaping their content knowledge.
from a classmate or an AI.” They described AI’s ability to differentiate
instructional materials, including translation support
- Secondary CS Teacher, Kentucky, USA for multilingual learners.
They also noted that AI’s ability to provide
Teachers also championed the “programmer’s mindset”
personalized feedback and additional examples
and shared about the value of computational thinking,
leads to a deeper understanding of information.
problem-solving, and creative skills developed in the CS
classroom.

The Role of AI in Computer Science Education: Results from a Teacher Survey 31

Guidance on the Future of Computer Science Education in an Age of AI

“It can be used to model or explain code and “AI can free up your time for more one-on-one
concepts to scaffold a student's learning. It can instruction by grading repetitive tasks and
create additional examples for students to providing instant feedback on student code.”
explain, create faulty code for ‘find the errors,’
- Secondary CS Teacher, California, USA
recommend sources for research, create
sample questions on a topic to help test-prep.”
By introducing students to AI early — how it works and
- Secondary CS Teacher, New York, USA how to use it — teachers prepare students for their
future personal and professional lives in the digital age.
CS teachers praised AI’s ability to automate tasks, such
as designing assessments and lessons and providing “[We are] preparing students for the future
feedback on student work. With support from AI in their where AI is a tool they can use and work with,
learning management systems and everyday classroom not against or replaced by.”
procedures, teachers can focus more on the students in
- Secondary CS Teacher, Maine, USA
front of them.

Ethical and Practical Considerations

What are the potential risks of using AI in the “Without supervision or intentional usage, it
could be misused to create code that students
CS classroom?
claim as their own.”

The most common concerns were about overreliance on - Secondary CS Teacher, Kentucky, USA
AI, plagiarism, and biases or ethical concerns. CS
teachers were particularly concerned that students Teachers also expressed concerns that biases in AI
would use AI instead of learning the content, losing both algorithms might impact the classroom community.
programming skills and the collaborative experience.
“AI systems have proved to be biased and to
“Students will not have to ‘struggle’ with produce incorrect coding solutions at times. It
problems. They can use AI to help debug code could lead to incorrect or biased learning for
or outline a program. Students will use AI for the students, as well as discriminated grading if
that collaborative problem-solving rather than an AI tool is used for grading.”
engaging with peers or instructors.
Communication skills are really important for - Secondary CS Teacher, Arizona, USA
students to develop.”
Finally, teachers discussed the importance of student
- Secondary CS Teacher, Montana, USA privacy and protecting sensitive information. While they
agreed that AI might benefit students by providing
CS teachers also described that, without such personalized support, they emphasized the importance
understanding, students might be more inclined to use of caution and clearly understanding what data is being
AI to cheat on their assignments. collected or used and how.

TeachAI is led by Code.org, ETS, the International Society for Technology in Education,
Khan Academy, and the World Economic Forum.

TeachAI & CSTA (2024). Guidance on the Future of Computer Science Education in an Age of AI. Retrieved from teachai.org/cs.
About TeachAI

TeachAI brings together education leaders and technology experts to assist governments
and education authorities in transforming education through teaching with and about AI.
The initiative is led by Code.org, ETS, the International Society for Technology in
Education, Khan Academy, and the World Economic Forum, and advised by a diverse
group of 150+ organizations and governments. TeachAI’s goals include guiding policy,
increasing awareness, and building community and capacity.

About CSTA

The mission of the Computer Science Teachers Association (CSTA) is to empower, engage,
and advocate for K-12 computer science teachers worldwide. With more than 25,000
members, CSTA supports and promotes the teaching of computer science and other
computing disciplines.
Guidance on the Future of Computer Science Education in an Age of AI

Visit teachai.org/cs

TeachAI is led by Code.org, ETS, the International Society for Technology in Education,
Khan Academy, and the World Economic Forum.

TeachAI & CSTA (2024). Guidance on the Future of Computer Science Education in an Age of AI. Retrieved from teachai.org/cs.