Basic Concepts in AI

Basic concepts in AI

Description

Introduction

This content pack introduces the foundational vocabulary and concepts of Artificial Intelligence that STEM higher education teachers need before exploring more advanced applications later in this module. Rather than treating AI as an abstract or purely technical subject, this introduction grounds each concept in real-world and environmental sustainability examples, drawing on established academic and technical sources, so that teachers can build conceptual AI literacy in their students while situating that literacy within socio-scientific environmental decision-making. By the end of this subtopic, teachers should be able to define AI in plain language, use its basic terminology correctly, distinguish AI systems from non-AI systems and AI-generated content from human-generated content, and help students recognise situations in which AI tools might genuinely be useful.

Types of AI

Description

Introduction

This subtopic moves teachers and students from knowing what AI is to understanding that AI is not one technology but a family of approaches, each suited to different tasks. Building on the foundational concepts from Subtopic 1.1, this subtopic introduces four practical categories: predictive AI, which forecasts outcomes from historical and real-time data; generative AI, which creates new content such as text, images, or code; classification and recommendation systems, which sort data or suggest actions; and adaptive AI, which combines several of these techniques and adjusts its behaviour as new data arrives. Drawing on established classification frameworks and the technical literature on machine learning, the goal is for STEM higher education teachers to classify tools accurately, explain the differences to students, and select the right type of AI for a given teaching or green-decision task, rather than treating every AI product as interchangeable.

This distinction matters because expecting the wrong type of tool to do a job it was never built for, such as asking a chatbot to precisely forecast next month's energy demand, leads to both unrealistic expectations and poor decisions. The subtopic walks through how each category emerged, what data and trade-offs it requires, and how to evaluate a tool's maturity before trusting its output, always tying the discussion back to concrete green-decision examples like crop-yield forecasting, environmental report drafting, and wildlife-trade detection. By the end, teachers and students should be able to look at any AI tool and ask "what type of AI is this, and does that match what I actually need?", a habit of critical, tool-appropriate thinking that carries directly into the chatbot-focused subtopic that follows.

AI Chatbots in Educational Contexts

Description

Introduction

Building on an understanding of AI types, this subtopic focuses on the form of AI that teachers and students are most likely to interact with directly: conversational AI chatbots such as ChatGPT, Gemini, and Claude. These tools sit within the generative AI category introduced in the previous subtopic and are increasingly used in higher education for brainstorming, explanation, feedback, and content drafting, a trend documented in a growing body of research reviewed later in this subtopic (Kasneci et al., 2023). Because chatbots are so accessible and so fluent, they raise AI literacy and decision-making questions in a particularly immediate way: how they work, how to use them well, and how to judge when to trust what they say.

AI tools in education

Description

Introduction

Having examined AI chatbots specifically in the previous subtopic, this subtopic widens the lens to the full range of AI tools available to STEM higher education teachers and students. Where Subtopic 1.3 focused on conversational, chat-based generative AI, this subtopic maps the broader toolkit identified in the project's WP2-A3 Best Practices Report against the wider research literature on AI in higher education: tools for content creation and presentation, assessment and feedback, writing and research support, classroom engagement, and adaptive or personalised learning. A systematic review of published research on AI applications in higher education found that the great majority of studies focused narrowly on adaptive systems, intelligent tutoring, and assessment (Zawacki-Richter et al., 2019), which is one reason this subtopic deliberately widens the discussion to the fuller range of tool categories STEM HE teachers now encounter.