Non-Visual AI Productivity

Back to Lessons

Lesson 2: AI Fundamentals in Simple Terms (No Vision Required)

Course: Foundations of Non-Visual AI Productivity (AI Basics)

Lesson content

• Lesson Objective: By the end of this lesson, you will be able to explain the relationship between AI, machine learning, and deep learning in simple terms, and understand why these basics matter for non-visual AI productivity.
• What is AI? Artificial Intelligence (AI) is software that learns patterns from data, then uses those patterns to make predictions or generate outputs. In plain language: AI learns from examples during training, then applies what it learned during use. Most AI models do not automatically “improve themselves” every time you use them; improvements usually come from newer versions, retraining, or fine-tuning by developers (and from you giving clearer prompts and better context).
• Analogy - AI vs traditional programming: Traditional software is like a recipe: a programmer writes exact step-by-step rules (“if X happens, do Y”). AI is more like a chef who has experienced thousands of dishes: given ingredients (input), it produces an output based on learned patterns, not a fixed recipe. For example, instead of hard-coding “cats have whiskers,” you show the system many examples and it learns what typically separates cats from dogs, similar to how people learn from exposure and practice.
• Machine Learning (ML): ML is the main technique behind modern AI. Developers train a model using lots of examples (text, images, audio, numbers) and adjust it so it becomes better at a task, such as predicting the next word, classifying an image, or grouping similar items. After training, the model can handle new inputs it hasn’t seen before by applying statistical patterns it learned, like a student who can solve new problems after enough practice.
• Deep Learning and neural networks: A common type of ML uses neural networks, and when there are many layers it’s called deep learning. Without getting technical, these are mathematical models that are very good at complex pattern recognition. With enough training data, deep learning can recognize speech, detect objects in images, and generate text or images by learning multiple levels of patterns. In 2026, Deep Learning is what allows AI to accurately describe your physical surroundings or 'read' the layout of a complex website for you.
• The Nesting Concept: AI is the big umbrella. Inside AI is machine learning. Inside machine learning is deep learning, which often uses neural networks with many layers. For low-vision or keyboard-first users who use some sight, this nesting helps you see how these terms fit inside each other when you hear them in tools, trainings, and feature lists.
• AI isn’t magic (and it can be wrong): AI doesn’t “understand” like a human. Many text AIs are excellent at predicting plausible language, which can produce impressive answers, yet they can also make confident mistakes. That’s why NVAIP treats accuracy as a workflow issue: you guide the model with clear inputs and you verify key outputs before you trust or share them.
• NVAIP connection - the workflow loop: In this course, we use one repeatable loop: Orient (clarify task and constraints), Plan (steps and checkpoints), Execute (work in small chunks), Verify (spot-check critical facts/numbers), and Recover (fix fast and keep versions). Even when learning theory (like today), we keep the same habits, especially Orient and Verify.
• Why non-visual interaction often works well (with caveats): Many AI tools are text-first (text-in, text-out), so you can access most of their power through typing or voice. However, some tasks are naturally visual (charts, layouts, dashboards). In those cases, NVAIP helps by requesting structured summaries (tables, bullet lists, “top 5 findings”) and by verifying important details, especially names, dates, and numbers.
• Mini exercise (3 - 5 minutes):
  1. Ask AI: “Explain the difference between AI, machine learning, and deep learning in 5 short bullets.”
  2. Then ask: “Give 2 common ways AI can be wrong and 2 ways to verify the answer.”
  3. Finally, verify one point by checking a trusted reference (for example, the IBM links in the References section of this lesson).

Continue

All lessons in this course

• Lesson 1
• Lesson 2
• Lesson 3
• Lesson 4
• Lesson 5
• Lesson 6
• Lesson 7
• Lesson 8

References

1. 1. IBM, “What is artificial intelligence (AI)?” IBM Think. https://www.ibm.com/think/topics/artificial-intelligence
2. 2. IBM, “AI vs. machine learning vs. deep learning vs. neural networks: What’s the difference?” IBM Think. https://www.ibm.com/think/topics/ai-vs-machine-learning-vs-deep-learning-vs-neural-networks
3. 3. IBM, “What Is Deep Learning?” IBM Think. https://www.ibm.com/think/topics/deep-learning