AI Glossary

AI (Artificial Intelligence): The simulation of human intelligence in machines that can perform tasks and learn from experience.

Prompt: A statement or question given to an AI model to guide its response or behaviour.

Prompt Engineering: The process of designing, refining, and optimising prompts to elicit desired outputs from AI models.

Backwards Design: An approach to instructional design that starts with the desired learning outcomes and works backwards to create activities and modules.

Project-Based Learning: An educational approach where students learn through real-world projects, promoting critical thinking and problem-solving skills.

Outcome: The desired result or goal of a learning experience or project.

Module: A self-contained unit within a course that focuses on a specific topic or set of related topics.

Learning Objectives: Clear and measurable statements that describe what learners should be able to do or understand after completing a lesson or module.

Real World Examples: Instances or scenarios from everyday life that illustrate the application of concepts or principles.

Activities: Engaging tasks or exercises designed to reinforce learning and apply knowledge in practical ways.

Step-by-step Directions: Clear and sequential instructions that guide learners through an activity or process.

Video Script: A written document that outlines the content and structure of a video lesson, ensuring a clear and organised delivery.

Discussion Questions: Thought-provoking inquiries that prompt learners to engage in critical thinking, reflection, and discussion.

Expand Learning: Suggestions or opportunities for learners to explore additional resources or extend their understanding beyond the core material.

Complexity: The level of intricacy, difficulty, or sophistication in the content or tasks presented to learners.

Sentence Variety: The use of different sentence structures, lengths, and styles to add variety and interest to written or spoken content.

Burstiness: A mix of short and long sentences that adds rhythm and emphasis to written or spoken communication.

Iterative Refinement: A process of continuously improving or adjusting prompts based on feedback and evaluation.

Multi-turn Prompts: Prompts that involve multiple interactions or exchanges with an AI model, creating a conversational experience.

Conversational Depth: The level of engagement and back-and-forth interaction achieved through multi-turn prompts.

AI Response Analysis: The examination and evaluation of the outputs generated by an AI model in response to different prompts.

Safe Outputs: AI responses that adhere to ethical guidelines and do not pose harm or risk to users or society.

Domain-Specific Prompts: Prompts tailored to specific industries, fields, or subject areas to generate contextually relevant outputs.

Creative Prompts: Prompts designed to stimulate imaginative and artistic AI outputs, such as storytelling, poetry, or music.

Ethical Considerations: The ethical principles and implications that arise in the design and use of AI prompts.

AI Misuse: The harmful or unintended consequences that can occur when AI prompts are used inappropriately or maliciously.

Privacy: The protection of personal information and data in AI interactions, ensuring user confidentiality.

Fairness: Ensuring that AI prompts and responses are free from bias and treat all individuals equitably.

Accountability: The responsibility and ownership of the outcomes and impact of AI prompts and their resulting outputs.

Societal Impact: The broader influence and consequences of AI prompts on society, including social, economic, and cultural aspects.

Project Guidelines: Specific instructions and requirements for the capstone project, providing a framework for implementation.

Review: A comprehensive assessment or recap of the course material, reinforcing key concepts and summarising the learning journey.

Capstone Project: The final project that integrates the knowledge and skills acquired throughout the course, applying prompt engineering to a real-world scenario.

Roadmap: A visual or written guide that outlines the sequential steps and milestones of a project or learning journey.

Expert Project Manager: An individual who oversees the planning, execution, and delivery of a project, ensuring its success and adherence to timelines.

Actionable Advice: Practical recommendations or suggestions that can be implemented to achieve specific goals or outcomes.

Project Plan: A detailed document that outlines the objectives, tasks, timelines, and resources required for a project's successful execution.

Evaluation Methods: Techniques and criteria used to assess the effectiveness, quality, or impact of prompts and AI responses.

Documentation: The process of recording and maintaining records of project activities, observations, insights, and results.

Iterative Process: A cyclical approach where improvements and adjustments are made through repeated cycles of evaluation and refinement.

Bias: Prejudice or favoritism that can be unintentionally introduced into AI prompts, leading to biased outputs or discriminatory behaviour.

Inclusivity: Ensuring that AI prompts and responses are accessible and inclusive for individuals from diverse backgrounds and abilities.

Iteration: Each cycle or round of refining and improving prompts and AI responses based on feedback and evaluation.

Adaptation: The ability to adjust and modify prompts and strategies based on new information, changes in circumstances, or evolving requirements.

Core Material: The fundamental content and concepts covered in the course, forming the foundation for learning and skill development.

Continuous Learning: The ongoing pursuit of knowledge and skill development beyond the scope of the course, staying updated on advancements in prompt engineering.

Prompt Solution: The complete set of prompts designed to address a specific problem or achieve a desired outcome.

User Experience: The overall experience and satisfaction of users when interacting with AI prompts and the resulting outputs.

Best Practices: Established methods or approaches that have been proven to be effective and produce high-quality results in prompt engineering.

Prompt Engineer: A professional specialising in the design, development, and refinement of prompts for AI models.

Prompt Crafting: The skill of carefully formulating prompts to elicit specific responses or guide the behaviour of AI models.

Zero-shot Prompting: Direct prompting (also known as Zero-shot) is the simplest type of prompt. It provides no examples to the model, just the instruction. You can also phrase the instruction as a question or give the model a "role," as seen in the second example below.

Chain-of-thought Prompting (C-o-T): In chain-of-thought prompting, the input question is followed by a series of intermediate natural language reasoning steps that lead to the final answer. Think of this as breaking down a complicated task into bite-sized, logical chunks.

Tree of Thought Prompting (T-o-T): The effectiveness of CoT prompting comes from its ability to break a complex problem solution into a sequence of smaller and simpler steps. Tree of thoughts (ToT) prompting similarly breaks a problem into a sequence of smaller steps—or thoughts—that are solved individually. However, the approach does not constrain the model to output these steps all at once. Rather, each thought is generated or solved independently and passed to the next step for solving the problem, which allows the model to:

• Explore multiple choices for each problem-solving thought.
• Evaluate whether certain thoughts bring the model closer to a final solution.
• Perform backtracking when certain thoughts are found to be a dead end.
• Search over a combinatorial space of possible problem-solving steps to find the best final solution.