1 of 42

Artificial Intelligence in Education

Hongjin Du, Megan Henderson, Xin Bai

​

York College of the City University of New York

2 of 42

  • Project Name: Artificial Intelligence in Education, a 2-part course module
  • Project Description: The CUNY team designed a 2-part course module for pre-service and in-service teachers to learn about AI and its educational implications. It includes engaging activities, tools the teachers can customize and adapt to their K-12 teaching, and valuable multimedia resources for further exploration. The slideshow will be used in two consecutive class sessions to introduce the critical AI concepts and educational tools and provide group activities to promote student engagement and inclusion.
  • Subject Area: Teacher Education
  • Grade Level: Undergraduate for preservice and in-service full-time DOE teachers�
  • Standards: We adopt the ISTE Standards for Educators in designing our course module.

(1) Standard 1—Learner

Educators continually improve their practice by learning from and with others and exploring proven and promising practices that leverage technology to improve student learning.

(2) Standard 4—Collaborator

Educators dedicate time to collaborate with both colleagues and students to improve practice, discover and share resources and ideas, and solve problems.

(3) Standard 5—Designer

Educators design authentic, learner-driven activities and environments that recognize and accommodate learner variability.

(4) Standard 6—Facilitator

Educators facilitate learning with technology to support student achievement of the ISTE Standards for Students.

  • Learning Goals:

(1) Comprehend fundamental artificial intelligence concepts including how it works and its latest applications.

(2) Recognize how distinct artificial intelligence concepts, tools, and implementations can foster student-centered learning experiences and bolster interdisciplinary objectives in teaching and learning.

(3) Contemplate strategies for utilizing artificial intelligence applications to enhance student performance, encourage students' curiosity and aptitude in computer science, and foster the development of STEM abilities.

(4) Evaluate ethical considerations and potential societal implications related to the implementation of artificial intelligence in education, promoting responsible and inclusive use of AI technologies in the classroom.

3 of 42

AI (Session 1)

4 of 42

��Time to sharpen our rusty drawing skills

https://quickdraw.withgoogle.com

  1. What did you like?
  2. What was surprising about the app?
  3. How is this an example of Machine Learning?
  4. What are some applications of such technology?

https://youtu.be/X8v1GWzZYJ4?list=RDCMUC_x5XG1OV2P6uZZ5FSM9Ttw

Game Time:

5 of 42

Game Time: What do these cats have in common?

6 of 42

What do these pictures have in common?

7 of 42

What is AI

A Working Definition

“The ability of machines to perform tasks that would normally require human intelligence–perception, reasoning, learning, and decision-making”

What do you already know about AI around you?

8 of 42

Rule-based Expert Systems

https://kunal3836.medium.com/expert-systems-and-machine-learning-3c130bf5d45d

9 of 42

Example: Teachable Agent �– a Rule-Based Expert System

10 of 42

11 of 42

The Knowledge Base: Facts and Rules

12 of 42

Machine Learning

  • A machine learning model is a mathematical representation of a dataset developed by AI, which can be used to support AI reasoning and problem-solving.

​

  • An example of ML (0:41-4:35):

https://www.youtube.com/watch?v=3bJ7RChxMWQ&t=6s#t=0m41s

​

13 of 42

What is a Neural Network?

https://www.youtube-nocookie.com/embed/bfmFfD2RIcg?playlist=bfmFfD2RIcg&autoplay=1&iv_load_policy=3&loop=1&modestbranding=1&start=

14 of 42

ANN – Let’s Tinker

https://playground.tensorflow.org

15 of 42

Teachable Machine by Google�- a video demo

16 of 42

Instructions for the Teachable Machine Groupwork

In this activity, you will use Google's Teachable Machine to train your own machine-learning model.

  1. Go to the Teachable Machine website and click "Get Started" to begin.
  2. New Project: Select what you would like to make an image project.
  3. Gather Training Data: Teachable Machine uses supervised machine learning, meaning you need to provide sample data for each class (i.e. category) that you create. Create two or more classes, give each class a label, and then add sample data to each class. The better the quantity and quality of your data, the better your model will work.
  4. Train: Once you have added in all of your sample data, click "Train Model."
  5. Test: Test your model in the preview window to see how well it works. The bars at the bottom indicate what class the AI predicts your test data falls into. If it doesn't work correctly all of the time, improve your training data and train your model again.
  6. Reflect: After you have trained your model, think: What surprised you about this process? How might you use a tool like this with K-12 students?

17 of 42

Some specific examples of how Teachable Machine can be used by K-12 students to develop critical thinking and problem-solving skills (from ChatGPT)

Identify waste for recycling: Students can train Teachable Machine to recognize different types of waste such as plastics, glass, and paper. The model could then be used in a waste sorting station to help identify and sort recyclable materials.

Assistive technology: Students can use Teachable Machine to train a model to recognize specific gestures or facial expressions that could be used to control a computer mouse, or as a communication aid for those with limited mobility.

Plant recognition: Students can train a machine learning model to recognize different types of plants, leaves, or flowers. The model could be used to help gardeners or botanists identify different plant species in the field.

Health monitoring: Students can use Teachable Machine to build a model that can detect changes in heart rate, breathing patterns, or other vital signs. The model could then be used to monitor patient health remotely, or as a tool for athletes to track their fitness levels.

Music composition: Students can train Teachable Machine to recognize different sounds and pitches, and then use this knowledge to create a model that can compose music based on the input. This can help students develop creative and artistic skills while also learning about machine learning algorithms.

18 of 42

Two Most Widely Used Generative AI Models

  • Generative Adversarial Networks or GANs: �technologies that can create visual and multimedia artifacts from both imagery and textual input data.

​

  • Transformer-based models — technologies such as Generative Pre-Trained (GPT) language models that can use information gathered on the Internet to create textual content from website articles to press releases to whitepapers.

​

19 of 42

How Generative AI Works

Photo credits: https://forums.stevehoffman.tv/threads/the-critic-cartoon-series.845377/

​

20 of 42

Selected tools for generating videos, images, or audio

21 of 42

Lab Time: AI via Scratch

Scratch with AI extension:�https://playground.raise.mit.edu/main/��

  1. Use an AI model trained through Teachable Machine: �Load in the scratch file below via the File dropdown menu: https://www.dropbox.com/s/ak7whll40htfwcn/Frog_EDUC231_AI.sb3?dl=0�
  2. Use other built-in trained AI models such as:
    • Face Sensing
    • Hand Sensing

​

22 of 42

How can generative AI tools be integrated into the curriculum so that students can learn about AI while also developing their own original digital media?

  • Teach how it works
  • Provide hands-on experience with generative AI tools: ChatGPT, StyleGAN, OpenAI Codex
  • Encourage trying new things and exploring possibilities
  • Use in the curriculum as a personal math tutor
  • Help create writing prompts and creative work.
  • Discuss the ethical and social implications

23 of 42

Readings:

​

​

24 of 42

Resources

Hands-On AI Projects for the Classroom (developed by ISTE and General Motors)

​

25 of 42

26 of 42

Exit Slips

https://forms.office.com/r/j3ZdpprFf4

27 of 42

AI (Session 2)

28 of 42

Ethics and AI

  • What is ethics?

The study of the moral principles and systems that govern a person’s behavior. It is the study of the systems people use to determine what is right and wrong.

​

  • AI-related ethical considerations:
  • All AI technologies are developed by humans.
  • Whether they have been preprogrammed with a set of rules, or use training data to learn, they will have biases based on human input and decision-making.
  • It is important that students understand that AI decisions are not objective, as well as to understand which stakeholders might benefit from certain biases in the technologies.

https://cdn.iste.org/www-root/2021-10/AI%20Ethics%20Guide%20EN.pdf

​

29 of 42

The Ethical Dilemma of Self-driving Cars

https://www.ted.com/talks/patrick_lin_the_ethical_dilemma_of_self_driving_cars?utm_campaign=tedspread&utm_medium=referral&utm_source=tedcomshare

​

30 of 42

31 of 42

A Guide on Ethics and AI

https://cdn.iste.org/www-root/2021-10/AI%20Ethics%20Guide%20EN.pdf

32 of 42

How does AI affect me and my students?

​

“It changes the skills people need to be employed. AI changes teaching, yes, but more important than that, AI changes the goals and purposes of teaching.”

–Christopher Dede, professor at the Harvard Graduate School of Education

33 of 42

ChatGPT

Teachers could use ChatGPT to:

  • design lesson plans
  • grade essays
  • develop rubrics
  • brainstorm instructions/prompts for assignments
  • write code

​

  • Have you used ChatGPT?
  • How comfortable are you with chatGPT?
  • The potential of chatGPT in schools?
  • The concerns of chatGPT in schools?

​

34 of 42

What are the implications of ChatGPT for Teaching and Learning

Engage critically and ethically

Empower students

Reimagine Pedagogy

35 of 42

Ethical Issues Surrounding AI

Algorithmic bias in commonly used tools like facial recognition

https://youtu.be/TWWsW1w-BVo

36 of 42

Some Ethical Principles Considered by Companies

​

  1. Fairness
  2. Reliability and Safety
  3. Privacy and Security
  4. Inclusivity
  5. Transparency
  6. Accountability

​

https://news.microsoft.com/cloudforgood/_media/downloads/the-future-computed-english.pdf

​

  1. Be socially beneficial.
  2. Avoid creating or reinforcing unfair bias.
  3. Be built and tested for safety.
  4. Be accountable to people.
  5. Incorporate privacy design principles.
  6. Uphold high standards or scientific excellence.
  7. Be made available for uses that accord with these principles.

​

https://blog.google/technology/ai/ai-principles/

37 of 42

38 of 42

Key Terms

  • Expert systems: An early AI computer system designed similar to the decision-making ability of a human expert.
  • Machine Learning (ML): The process of giving a computer a sample set of data, so that it can “learn” by making associations between data and draw conclusions without the need to program explicit instructions.
  • Generative AI: AI that uses machine learning algorithms that enable machines to create fake data or media - such as video, images, text, and audio - based on the features of its training data
  • Deep Learning: A subset of machine learning in which multilayered neural networks learn from vast amounts of data.
  • Ethics: The study of the moral principles and systems that govern a person’s behavior. It is the study of the systems people use to determine what is right and wrong.

39 of 42

Readings:

  • Wired Magazine Article on Perception�In the article, Reality is not what it seems: the science behind why optical illusions mess with our minds, the science, and mystery behind perception is explored.�
  • Marl/O—Machine Learning for Video Games�A video that shows machine learning and neural networks in a video game—great for students!

40 of 42

Resources

​

41 of 42

42 of 42

Exit Slips

https://forms.office.com/r/j3ZdpprFf4