brightlab.Made for curious minds
High school · 2 lessons

Can you trust it?

AI can answer questions, write essays, make pictures and copy voices. Before you trust it, check it: is this AI any good, and is this picture, video or voice real? Watch short animations, make a prediction, then test the idea yourself in a hands-on lab.

Start with evals For teachers

Is this AI any good?

  1. Who chose the examples? A demo shows the best someone found.
  2. Could it have seen the test? Fresh questions beat famous ones.
  3. How many questions? A handful of results is mostly luck.

Is this picture, video or voice real?

  1. Who posted it? Investigate the source before the pixels.
  2. Who else is reporting it? Real events leave more than one trace.
  3. Where did it first appear? Trace it back to the original.

The lessons

Each one stands on its own and takes about an hour.

01 · Evals · Year 9

Put the bots to the test

Two homework bots both say they’re brilliant. How would you find out which one really is?

3 animations · Bot bake-off · 60 min
02 · Deepfakes · Years 7–12

Real or fake? Check before you share

A video looks completely real. How can you tell whether to believe it?

3 animations · Check it like a pro · 60 min

How each lesson works

The same rhythm as every Brightlab investigation.

  1. 01

    Watch

    Short silent animations show the idea in action.

  2. 02

    Predict

    Commit to what you think will happen, and why.

  3. 03

    Test

    Try it in the lab and gather evidence.

  4. 04

    Explain

    Use the evidence. Hunt for a counterexample.

For teachers

Each lesson page ends with a sequence, assessment evidence, common misconceptions and a glossary.

Who each lesson is for

  • Evals (Year 9): Fractions, percentages and averages. No coding or AI knowledge needed.
  • Deepfakes (Years 7–12): Percentages. Years 7–8 can focus on the verification desk; Years 10–12 can go further with the detector maths.

What is included

  • Animations under 40 seconds, with captions and a written description of every step.
  • A browser lab for every lesson. Nothing is sent to a server, and nothing is stored after the page closes.
  • Explanation prompts and a counterexample challenge for each lesson.

The animations were made with Manim. In the evals lesson, Blitz and Sage are pretend bots and Brightlab wrote all of their answers; the animations and the lab score the same question file, so the numbers match. In the deepfakes lesson every post, account and person is invented, and no real deepfake is shown.