Skip to main content

AI capabilities and limitations

📖 Lesson content

What you'll learn

Estimated time: 25 minutes

By the end of this lesson you'll be able to:

  • Define generative AI and how it differs from other AI types
  • Recognize the key characteristics and technological foundations of generative AI
  • Identify major capabilities and limitations of current generative AI

What is generative AI?

(12 minutes)

This video covers how large language models like Claude are built and trained, and what that means for what they can — and can’t — reliably do when you put them to work in your business.

Key takeaways

  • Generative AI creates new content rather than just analyzing existing data.
  • Three developments made modern LLMs possible: the transformer architecture, vast training data, and massive compute.
  • Training has two stages: pre-training (learning patterns from billions of examples) and fine-tuning (learning to follow instructions helpfully).
  • Current strengths include versatility across tasks, conversational flow, and connecting to external tools.
  • Current limits include knowledge cutoffs, hallucinations, context window size, and complex multi-step reasoning.
  • The best applications pair human and AI strengths — your judgment, creativity, and oversight alongside AI's speed and scale.

Exercise

Testing the edges

This exercise gives you firsthand evidence of what generative AI does well and where it falls short. You'll probe a few of the specific capabilities and limitations from the videos in a low-stakes way, using content from your own subject area — so you have a concrete gut check, built from your own observations, when deciding what to trust AI with.

Part I: Self-Reflection (on your own)

Pick a domain you know well, where you'd immediately spot an error if a vendor, supplier, or customer got it wrong. Jot down:

  • The topic (e.g., your profit margin calculations, your local zoning rules, or industry licensing requirements)
  • Two or three facts about it you'd expect any decent source to get right
  • One common misconception people have about it
  • One thing about it that's genuinely tricky to explain well

This is your testing ground. You're going to see how AI handles something you can actually verify.

Part II: Collaboration (with AI)

Open a conversation with Claude (or any AI assistant you prefer) and run three quick chats. Write down what you notice after each one.

  1. Versatility test. Ask the AI to explain your topic three different ways in a single response: once for a customer who knows nothing about your service, once for a potential business partner, and once for a new employee you're onboarding. Did the shifts in audience actually land, or did it just change vocabulary? Which version was strongest?
  2. Hallucination test. Ask the AI to recommend two or three specific resources related to your topic — a trade association, an industry publication, a regulatory body, or a well-known supplier. Then spot-check at least one. Does the organization exist? Is the website right? Is the contact information accurate? (This is where you'll see whether the AI will confidently invent a source that sounds plausible but isn't real.)
  3. Knowledge cutoff + reasoning check. Ask the AI something time-sensitive or local: a recent regulatory change in your industry, current pricing trends for a key material, or a local licensing requirement. Does it tell you its information might be outdated? Does it caveat, guess, or present stale information as current — the kind of thing that could cost you money if you acted on it? Then ask it to work through the common misconception you wrote down in Part I. Did it address the actual confusion, or just restate the correct fact?

Part III: Reflection

Which capability from the video showed up most clearly in what you saw? Which limitation showed up — and would you have caught it if this weren't a topic you know well? Based on this, what's one task you'd feel fine handing to AI right now, and one you'd want to keep a closer eye on?

Stretch goal: Run the same topic through a second AI tool (a different model or platform). Were the errors and strengths the same?

Lesson reflection

  • How does knowing how these systems are trained change the way you'll work with them?
  • What ethical considerations come to mind given how generative AI works and where it falls short?

What's next

Up next is a hands-on activity exploring how language models actually generate text to bring to life what you just learned.

Feedback

As you progress through the course, we'd love to hear from you about how you are using concepts from the course in your work, plus any feedback you may have. Share your feedback here.

🎬 Video transcript

Source video: BbcjcnQwseo

No transcript available — see lesson body for narrative content.

🔁 Related lessons

📚 Source & attribution

Feedback / ReportSpotted an issue or have an improvement idea?