AI & Cybersecurity for Teens

Table of Contents

• Activity 0: Intro to NetsBlox, RPCs and Simple Chat App
• Activity 1: Cryptography & Coding a Caesar Cipher
• Activity 2: Intro to Chat App
• Activity 3: Intro to Network Security/Micro:Bit Activity
• Activity 4a: Client-Server Chat App
• Activity 4b: Denial of Service Attacks (Extending Chat App)
• Activity 5a: Sentimental Writer App
• Activity 5b: Intro to Sentiment Analysis & Cyberbullying
• Activity 6: Intro to Machine Learning & Classification
• Activity 7: Data Exploration with CODAP
• Activity 8: Twitter Bot Detection
• Activity 9: Registration Bots and Gradient Descent
• Activity 10: Adversarial Examples
• Activity 11: Generative Adversarial Networks (GANs)

Activity 0: Intro to NetsBlox, RPCs and Simple Chat App

Estimated Duration	2 periods (45 minutes)
Lesson Summary	In this lesson, students will be introduced to NetsBlox, specifically the messaging blocks to create a ‘simple chat app’ (which will be extended in subsequent projects).
Learning Objectives	Students will: learn about NetsBlox, and its key abstractions for RPCs (Call block) and messaging (Send/Receive Message).
Vocabulary	Distributed Computing; Remote Procedure Calls; Message Passing
Activities	Create your NetsBlox account Code-along the Map & Covid App (first 50 minutes of this video) Simple Chat App (Starts at 40 min mark))
Student Resources	Intro Project for NetsBlox RPCs
Teacher Resources	Service Docs: Google Maps Service Docs: Weather Service Docs: Geolocation
Additional Resources	Explore netsblox.org Intro to ACT Slides (use as needed to orient students to "Why AI+Cybersecurity")

Activity 1: Cryptography & Coding a Caesar Cipher

Estimated Duration	1 class period (45 minutes)
Lesson Summary	In this lesson, students will look at the historical significance of the Caesar Cipher and learn how to implement the shift-cipher using both paper/pencil and NetsBlox.
Learning Objectives	Students will: learn about a simple shift-cipher encrypt/decrypt messages using paper and pencil implement shift cipher in NetsBlox
Vocabulary	Cipher; Encrypt; Decrypt; Plaintext; Ciphertext
Cybersecurity Concepts/Topics	Encryption; Obfuscation of information
Activities	Introduction to encryption & historical context of Caesar Cipher Paper/Pencil implementation of Caesar Cipher Netsblox implementation of Caesar Shift Cipher
Real-World "Hook"/Issue	Data Encryption & Privacy
Discussion Questions	What data needs to be encrypted? What are the downsides of encrypting data? How many possible keys are possible with a Caesar Shift Cipher?
Unplugged/Other Pre-activity	Print the Caesar Cipher wheel on cardstock (requires brad fasteners) Print Caesar Cipher Worksheet
Student Resources	Caesar Cipher Activity in NetsBlox Brute Force Cracking Caesar Cipher
Teacher Resources	NetsBlox Cryptography Exercises Caesar Cipher (Wikipedia) Practical Cryptography: Caesar Cipher
Additional Resources	CryptoClub: Desert Oasis (Game) TED Talk: Why Privacy Matters Time it takes to brute force a password (aka why we need special characters)

Activity 2: Intro to Chat App

Estimated Duration	2 class periods (45-50 minutes each). One for chat app and one for encryption/decryption of messages.
Lesson Summary	In this lesson, students are introduced to the message passing functionality of NetsBlox to create a simple chat app. (The app is extended in subsequent projects to address additional topics in this curriculum).
Learning Objectives	Students will: Be introduced to the message creation and message send/receive blocks in NetsBlox Create a simple message passing application in NetsBlox Encrypt/decrypt messages sent to each other
Vocabulary	Message passing; Encrypt; Decrypt; Address; Role
Cybersecurity Concepts/Topics	Encryption
Activities	Code-along a Simple Chat App (starts at 40 minute mark) Incorporate encryption functionality from the Caesar Cipher activity to encrypt/decrypt messages sent to each other
Real-World "Hook"/Issue	Messaging is a big part of teen communication. What are the pros/cons? Why are messages in apps like WhatsApp encrypted? Discuss this article: What are the ethical considerations?
Discussion Questions	What are some ways in which you communicate with friends? Which is most common? Would you like to keep these chats private and confidential? What are some situations in which such chat should be monitored? What is the fundamental idea behind apps like chat or text?
Student Resources	Activity Walkthrough
Teacher Resources	Adding Encryption to the Chat App

Activity 3: Intro to Network Security/Micro:Bit Activity

Estimated Duration	1 class period (45-50 minutes)
Lesson Summary	In this lesson, students will learn about port scanning and network security.
Learning Objectives	Students will: learn about scanning for devices on a network explore port scanning
Vocabulary	Port Scanning - looking at every possible entry-point into a computer; Firewall - Software tool used to monitor and block ports; Protocol - Set of rules that defines how communication would be sent
Activities	Students will need a Micro:bit & Computer. There should also be 2-4 Micro:bits used as “servers” in the class. Micro:bit - Server Micro:bit - Port Scanner
Student Resources	Activity Instructions
Teacher Resources	Nmap (Wikipedia) Micro:Bit Website

Activity 4a: Client-Server Chat App

Estimated Duration	1 class period
Lesson Summary	In this lesson, students are introduced to the need and role for a server to moderate and mediate communication via messages. This means we need to create a server app and have client apps communicate through the server.
Learning Objectives	Students will: Be introduced to a client-server version of the chat app Understand how a chatroom works and can be created in NetsBlox Understand the role of a server app- to 1) accept connections from client apps, 2) receive their text messages and 3) resend them to every other client. Each user who wants to participate in the Chatroom needs to run a client that 1) connects to the server, 2) sends text messages to it, 3) accepts messages from the server, and 4) displays them on its screen.
Vocabulary	Client; Server; Global address; Local address
Cybersecurity Concepts/Topics	Message passing; Content moderation; Encryption
Activities	Code-along a client-server chat app Extend it to have the server play other roles such as message translation, trying to break messages encrypted using a Caesar Cipher
Discussion Questions	How is this chat app different from the first one? Why did we change it? Why do we need to connect to the server? Can we still fool the server so it can't break our encryption?
Student Resources	Chatroom Activity Walkthrough
Teacher Resources	Completed Chat Client 🔒 Completed Chat Server 🔒
Possible Extensions	Translation Color-coding messages based on sender Persistence (show the last 10 messages on first connect) Scrolling
Post Activity Discussion/Reflection	How are secure messages passed over the open web? Public Key Encryption - CS Unplugged Activity & Computerphile Overview of PKE & Diffie-Hellman

Activity 4b: Denial of Service Attacks (Extending Chat App)

Estimated Duration	1 class period
Lesson Summary	In this lesson, students extend the previous chat application to explore denial of service (DoS) attacks.
Learning Objectives	Students will: Perform a denial of service attack by overwhelming it with messages Work together to add safeguards to the server to prevent DoS attacks Reflect on approaches and connections to classification and rule-based AI
Vocabulary	Denial of Service (DoS); Distributed Denial of Service (DDoS); rate limiting
AI/ML Concepts/Topics	Rule-based AI: An AI system that comprises a set of human-coded rules that result in pre-defined outcomes (similar to if-then logic).; Classification
Cybersecurity Concepts/Topics	Denial of Service; Distributed Denial of Service
Activities	Discuss Denial of Service attacks Discuss how DoS attack can be simulated in the client-server chat app (What will make the server unable to accomplish its role) Students simulate an actual attack on the Chat App server Discuss the ways in which the server can protect itself from such attacks Implement one or more such strategies in student groups End discussion with the idea of a rule-based AI system that can use data to develop an if-then set of rules to mitigate DoS attacks
Discussion Questions	Are there any weaknesses to the mitigation techniques you used? What if the user made up new usernames? Could you verify that the user is not a bot?
Student Resources	Starter project to code a technique for mitigating DoS attacks
Teacher Resources	Lesson Slides
Possible Extensions	Add a registration step for users so they have to create an account before sending messages Add a login step where the user needs to provide a password Use a token (such as a big random number) issued on successful login to ensure the user isn't spoofing their username
Post Activity Discussion/Reflection	Using examples from real-life, what are the impact/dangers of such attacks? Do you know of any activism-motivated DDoS attacks? Captchas are also one example of how this is mitigated in industry (and related to multi-factor authentication when attempting to login).

Activity 5a: Sentimental Writer App

Estimated Duration	2 periods
Lesson Summary	In this lesson, students will learn about classification, and the use of natural language programming and an openly available API (ParallelDots) to classify text.
Learning Objectives	Students will: learn about classification through AI/ML. explore classification in NetsBlox through the sentimental “writer” - a basic program that uses an API to classify content (i.e., words/music).
Vocabulary	Classification; Natural Language Processing; Sentiment Analysis; Polarity; API
Activities	Introduction to NLP and Sentiment Analysis Twitter Sentimental Writer (unplugged movies reviews) Music (song lyrics) Sentiment using Genius API (NetsBlox/View in Python)
Student Resources	Slides with guided activity description Sentimental Writer Starter Code
Teacher Resources	Sentiment Writer Teaching Guide 🔒 Sentiment Writer Slides 🔒 Instructional Video (~17 minutes) 🔒
Possible Extensions	Classify music lyrics/NYT articles
Additional Resources	How NLP works

Activity 5b: Intro to Sentiment Analysis & Cyberbullying

Estimated Duration	1 class period
Lesson Summary	In this lesson, students will extend the Client-Server Chat app and combine it with the Sentimental Writer to prevent unsavory words and phrases from being sent to others via chat.
Learning Objectives	Students will: discuss the issue of online safety and cyberbullying, the impact of harmful language explore how we can use AI to detect and prevent sending hurtful words via chat
AI/ML Concepts/Topics	Classification; Sentiment Analysis; Toxic/Abusive language detection; Natural Language Processing
Cybersecurity Concepts/Topics	Cyberbullying; Online safety
Activities	"Chat App meets Sentimental Writer" - extend the chat app to use the ParallelDots service to determine if we should flag/block text.
Real-World "Hook"/Issue	White Hat Hackers Stand Up to Cyberbullies TIME's First-Ever Kid of the Year
Discussion Questions	What are some of the harms teens and young people face on social media & the internet? Now that we know that we can use AI to classify text as negative or hate speech, or sarcasm, how can we use it to tackle the use of harmful language in cyberspace? How is cyberbullying related to cybersecurity? How do you think apps like “Safe Search” or automated censoring of content (posts/tweets) work?
Student Resources	Starter Project
Teacher Resources	Chat Server (only one needs to be run by the teacher) 🔒 Lesson Slides 🔒
Post Activity Discussion/Reflection	What can technology companies and society do to prevent cyberbullying? Tinder will now use AI to ask if you really want to send that offensive message Perspective API

Activity 6: Intro to Machine Learning & Classification

Estimated Duration	1 class period
Lesson Summary	In this lesson, students will learn about machine learning as an AI technique and what it means to train a model. They will explore AI applications using pre-trained models available online.
Learning Objectives	Students will: learn about machine learning as an AI technique learn about classification and explore what it means to “classify” something using machine learning models
Vocabulary	AI; Machine learning; Classification; ML Model; Training a model
Activities	Discuss how AI is all-pervasive today Watch ML video/slides (in student resources), discuss Explore one or more AI/ML tools and share Discuss rule-based AI vs ML
Real-World "Hook"/Issue	Any big AI/ML example from everyday life including ChatGPT, Netflix, Amazon, etc
Discussion Questions	How is rule-based AI different from machine learning?
Student Resources	Teachable Machine Drawing Completion Tool (SketchRNN) Recognizing emotions in pictures
Teacher Resources	Lesson Slides
Additional Resources	More machine learning exploration links & resources

Activity 7: Data Exploration with CODAP

Estimated Duration	1 class period
Lesson Summary	In this lesson, students will learn about data visualization and gain hands-on experience exploring various datasets with CODAP.
Learning Objectives	Students will: learn about data distributions learn how to explore data with CODAP learn about periodic data
Vocabulary	Distribution; Periodic
Activities	Complete CODAP Tutorial (parts 1 & 2) Explore the birthday dataset as a class (answer questions in link) Explore another dataset of interest (from the additional datasets link)!
Discussion Questions	Discussion questions can be found in the CODAP links
Student Resources	CODAP Tutorial Part 1 CODAP Tutorial Part 2 Birthday Dataset in CODAP Additional CODAP Datasets
Teacher Resources	CODAP Educator Resources

Activity 8: Twitter Bot Detection

Estimated Duration	1 period
Lesson Summary	In this lesson, students will create rules to classify (synthetic) Twitter accounts as bots or humans then learn how they can automate this process to learn a decision tree from data.
Learning Objectives	Students will: Gain hands-on experience with rule-based AI Learn how rules can be automatically selected to construct a decision tree from data
Vocabulary	Classify; Building a decision tree; Entropy; Features
AI/ML Concepts/Topics	Decision Tree; Rule-based AI; Entropy
Activities	Introduce bot accounts using a real-world hook Decision Tree Unplugged Activity Explore the Twitter dataset using CODAP & create rules to classify accounts Reflect on how we created our rules and how we might automate this Decision tree learning block as Parsons problem
Discussion Questions	Does changing the data change the learned tree? How could you make a bot to fool the tree? What features would it need to have? What bot would fool the model and tweet the most?
Unplugged/Other Pre-activity	Decision Tree Unplugged Activity
Student Resources	Starter Code for Decision Tree Activity
Teacher Resources	Solution to Decision Tree Activity 🔒
Additional Resources	How to spot fake social media accounts How to tell if you’re talking to a bot Spot the Troll (game)

Activity 9: Registration Bots and Gradient Descent

Estimated Duration	2 class periods
Lesson Summary	In this lesson, students will be introduced to optimization, a fundamental concept in machine learning. They will then use gradient descent to optimize a simple bot detection model.
Learning Objectives	Students will: Program a bot to register for an account Train an ML model using gradient descent
Vocabulary	Optimization; Gradients; Gradient Descent
Activities	Introduce registration website activity in NetsBlox Program a bot to register for an account Find the Minimum Game Implement a model for bot classification (Parsons problem)
Student Resources	Account Registration Activity in NetsBlox Starter Code for Registration Bot Find the Minimum! Starter Code for Bot Classification
Teacher Resources	Solution to Bot Classification Activity 🔒
Additional Resources	Visualizing Different Optimization Algorithms

Activity 10: Adversarial Examples

Estimated Duration	2 periods
Lesson Summary	In this lesson, students will learn how they can use what they learned about optimization to optimize points in a way to fool a model. They will also gain hands-on experience creating adversarial examples both in NetsBlox and using another online tool.
Learning Objectives	Students will: Apply the concepts they learned around optimization to optimize a point (to fool the model) rather than a model Implement an optimization algorithm for generating adversarial examples
Vocabulary	Adversarial examples; Optimization; Model parameters
AI/ML Concepts/Topics	Gradient descent; Optimization; Adversarial examples; Classification
Activities	Visualize the predictions of the model in the previous activity for many made-up points Discuss what it would mean to move a point toward a minimum or maximum wrt the model predictions Implement an optimization algorithm on a point with the objective of fooling the model Explore an online tool for generating adversarial examples (link below)
Discussion Questions	What implications could adversarial examples have for self-driving cars? What would it mean to generate adv. examples for a facial recognition model?
Student Resources	Break Neural Networks in the Browser
Teacher Resources	Completed Adversarial Example (check out the adversarial point sprite) 🔒

Activity 11: Generative Adversarial Networks (GANs)

Estimated Duration	2 class periods
Lesson Summary	In this lesson, students will learn about generative adversarial training as they explore the question "Rather than generating a single point, can we train a model to produce realistic looking points?"
Learning Objectives	Students will: Use the optimization concepts from a earlier lessons to brainstorm how we could optimize a model to make adversarial examples Implement the training algorithm for a generative model
Vocabulary	Generative-Adversarial Networks; Generator; Discriminator
Activities	Introduce the Circle GAN project (show the working example) Sketch out the pseudocode for training a model to produce adversarial examples (in the slides) Implement the training algorithm for the generator Train a GAN in the browser using GAN Lab
Discussion Questions	Think back to the Find the Minimum game. What would the y-axis represent for the generator? What would the x-axis represent? What would the y-axis represent for the discriminator? What would the x-axis represent?
Student Resources	Circle GAN Starter Code GAN Lab
Teacher Resources	Circle GAN Solution 🔒
Additional Resources	Train a GAN in Python