AI in STEM Education

• This curriculum explores Artificial Intelligence (AI) and its applications in education, particularly within STEM.
• It introduces the 5 Big Ideas in AI using the Internet of Things (IoT) in STEM education.
• The Al4K12 initiative proposed these ideas, focusing on Societal Impact.
• The module aims to help educators integrate AI concepts into elementary and secondary classrooms through examples, activities, and scenarios.

Key Notions and Concepts

• AI combines computer science and robust datasets for problem-solving.
• AI involves computer systems performing tasks requiring human intelligence like visual perception, speech recognition, decision-making, and language translation.
• A robot is an automatically operated machine that replaces human effort without necessarily resembling humans.
• Capacity-based AI is categorized into artificial narrow intelligence, artificial general intelligence, and artificial superintelligence, based on their learning and application capabilities.
• Functionality-based AI sorts AI by how it applies learning to process data, respond to stimuli, and interact within environments.
• The Al4K12 project focuses on teaching AI in K-12 education in the US, aligning with the CSTA standards for computer science education.
• IoT uses computer tools to automate real-world processes, reducing direct human participation.
• IoT is defined as the ability of computing devices to transfer information over a network without human-to-human or human-to-computer interaction.

Introduction to Artificial Intelligence (AI)

• AI can simulate human intelligence using programs and is used in various products and services.
• AI focuses on developing algorithms and systems for tasks requiring human intelligence, such as reasoning, self-learning, and decision-making.
• Not all AI is a robot but robots are a more complex form of AI.
• AI can learn in three ways: supervised, semi-supervised, and unsupervised.
• Supervised learning involves human guidance, like identifying images.
• Semi-supervised learning involves AI classifying new images based on a pre-existing dataset.
• Unsupervised learning involves AI identifying patterns and differentiating images autonomously.

Types of Artificial Intelligence

• Artificial Intelligence exists simultaneously in various types.
• Classifications of AI depend on the approach, such as Capacity-based, Intelligence-based, Functionality-Based etc.
• Capacity-based AI:
  • Artificial Narrow Intelligence (ANI) executes very specific actions and cannot learn beyond its design.
  • Artificial General Intelligence (AGI) can learn, think, and perform tasks like a human.
    • Creating machines that can multitask and act in everyday life is the goal.
    • Supercomputers and generative AI models, like ChatGPT are used.
  • Artificial Superintelligence (ASI) refers to machines that surpass human intelligence.

AI Functionality

• AI is sorted into four different types regarding functionality.
• Reactive Machines react immediately without storing memories or learning from past experiences.
• Limited Memory AI stores past data to make predictions, improving performance through learning.
• Theory of Mind AI picks up on emotions and subtle environment changes to act on them.
• Self-Aware AI reaches a singularity, becoming independent and beyond human control.

AI in the Classroom

• Teachers can introduce AI by showing images of AI applications, like robots and virtual assistants and then asking students.
• It is important to ask students questions to encourage them to think about AI and its applications.
• A suggestion is to create groups and separate the Als to discuss the different types of AI.

AI in Everyday Life

• AI is in people's routines, from facial recognition to Netflix recommendations.
• AI is in healthcare, finance, manufacturing, transportation, entertainment, and education.
• AI algorithms are used in search engines, social networks, and virtual assistants.
• AI is used in autonomous transportation, medical diagnoses, and adaptive educational platforms.
• Smart homes, thermostats, and safety systems use Al to learn user preferences.

5 Big Ideas of AI

• John Haugeland proposed the "Five Big Ideas" which summarize the key concepts that make up the area of AI.
• These ideas offer an overview of the fundamental principles behind AI development.
• The Five Big Ideas are:
• Perception: AI enables computers to ‘see' and ‘hear' via sensors, extracting meaning from sensory signals.
• Representation & Reasoning: AI maintains representations of the world and uses them for reasoning.
• Learning: AI generates new representations through machine learning, with inputs from humans or the machine.
• Natural Interaction: AI requires knowledge to interact with humans naturally, using human-like qualities.
• Societal Impact: Technology has impacted lifestyle but a balance is important.

Societal Impact in AI

• AI has a significant impact on society, from the economy to the ways people live their lives, transforming society.
• Al also presents benefits and challenges.
• AI is used to automate repetitive tasks, improve healthcare, advance transportation, personalize education, and enhance e-commerce.
• AI can improve accessibility, enhance cybersecurity, raise ethical issues, and aid government decision-making.
• Societal Impact of Al can be positive or negative because data biases during AI training can lead to skewing results.

Internet of Things (IoT) and AI

• IoT is strongly connected to AI
• IoT helps the AI to process and create an appropriate response to a real-world event signaled by the IoT .
• Al is essential to IoT.
• Complex real-world systems rely on AI to respond to "trigger switch is pressed, and it's dark, turn on light A.”
• AI can help the IoT to reduce the need for direct human participation.
• Mimicking human capabilities will increase AI contribution to IoT applications.

Where IoT and AI Interact

• Robots in Manufacturing: robots equipped with AI algorithms learn from new data provided by IoT sensors.
• Helps to save time and costs.
• Self-driving Cars: predict behavior using data on road conditions and weather and AI connected with the IoT.
• Cars can improve with each trip.
• Smart Thermostat Solutions: AI-powered IoT checks and manages temperature anywhere based on preferences.

Ethics in AI

• Ethical issues include fairness, data responsibility, privacy, climate degradation, human rights, accountability, and technology misuse.
• AI in education raises concerns about privacy, data access, and over-reliance on technology.
• AI-powered devices can collect data of behavior, preferences, and learning patterns.
• Over-reliance on AI may prevent students from developing social skills like communication, perseverance, or critical thinking.