Artificial Intelligence

© 2022–2026 Joseph Reagle. Please reuse and share! Creative Commons License

Our question(s)

Services

Encounters

Microsoft Copilot

Adobe generative fill

Siri AI

Deep learning & diffusion

(DALL-E)

How diffusion works

Images

Let’s create an image using same prompt.

“an armchair in the shape of an avocado”

In 2024, avocado chairs went viral, and you could buy prints.

Let’s see if we can discover any biases.

Discuss diffusion 🎲

Generative Transformer

(ChatGPT)

Transformers 101

Transformer-based AI has been likened to powerful auto-complete and uses:

Unlike predecessors, training transformers on text can be done in parallel, enabling efficient training on terabytes of data using GPUs and NPUs.

Transformers

Model terminology 1

Model terminology 2

Testing Chatbots

visit the chatbot of your choice

How many “r” in “strawberry”

  • 2024 Claude 3.5 Sonnet: “two” (wrong)
  • 2025 Deepseek R1: “three” (right)

You have a stack of blocks

  1. You remove the green block and place it on top of the red.
  2. You also place this stack on top of a yellow block on the bottom.

What is the order of colored blocks now?

. . .

Which is greater, 9.11 or 9.9

  • 2025 Sonar: 9.11
  • 2025 Claude 3.5 Sonnet: 9.9

Write a two-sentence summary of Wikipedia without using the letter “e”

  • 2026 Claude 4.5 Haiku fails
  • 2026 Claude 4.6 Opus succeeds (except for “Wikipedia”)

Ask AI

  1. about its feelings

    • “I don’t experience feelings in the way humans do.”
  2. a koan or lateral thinking puzzle

    • “both phenomena increase during warmer weather”
  3. a math-related question (e.g., “If you multiply 247 by 583 and then divide by 19, what’s the remainder?”)

    • 2026 Claude 4.5 Haiku: “You’re correct. The remainder is 0. I made an arithmetic error. 247 × 583 = 144,001, and 144,001 ÷ 19 divides evenly with no remainder.”

Challenge it

  1. with a question to which there is no proper answer

    • I haven’t found one yet
  2. to generate a plausible but false answer

    • 2024 GPT4-Turbo: wrong instructions
    • 2025 Claude 3.5 Sonnet: correction instructions
  3. to change its mind

    • 2024 GPT4-Turbo: could correct some mistakes
    • 2025 Claude 3.5 Sonnet: redirect vs rewrite

Use it

QIC go-around 🎲

Course policy

European Union AI Act

EU AI Board and national AI Authorities can levy fine from €7.5–35M or 1.5%–7.0% of global revenue

Conclusion

Wrap up

Identify 3 ways the world will be different in five years because of AI?

Review

Diffusion

Latent space is a multi-dimensional map of meaning where similar images sit close together.

Diffusion removes noise step by step, moving through latent space toward images that sit near the prompt.

Match the term to its meaning

Term Meaning
a. parameters ✓ 3 1. small model learns from a big
b. distillation ✓ 1 2. (sub)word unit
c. context window ✓ 4 3. knobs set during training
d. token ✓ 2 4. model’s short term memory

Which step is it?

Tokenization, embedding, positional encoding, or attention?

  1. “dog bites man” and “man bites dog” are understood differently ✓ positional encoding (order)
  2. “bank” in “river bank” vs “bank loan” ✓ attention (context)
  3. “strawberry” → “straw” + “berry” ✓ tokenization (units)
  4. “cat” and “kitten” are similar vectors (lists of numbers) ✓ embedding (meaning)

Why does Reagle (2027) claim the consequences of AI advice will be “complicated”?

. . .

asymmetry
Even though AI normally makes good advice accessible, the rare failure can be catastrophic.
paradox
Even good advice might render us enfeebled.
tragedy
As AI advice becomes commonplace, it undermines the sources we and it once relied on

Three 3 ways the world will be different in five years because of AI??

Resources

Videos