Future Program Preview

Available later

Chatbot Character Lab

Design a fictional chatbot character, test whether it follows its rules, and collect the moments when fluent language hides a weak or wrong answer.

  • Prompts
  • Language
  • State
  • Testing
  • Safety
  • Human Judgment
Available laterAvailable later. No external chatbot account, child data submission, or current AI class is implied by this preview.

Future build

What the project is intended to contain.

  • An original fictional character brief
  • Behaviour rules
  • A test conversation set
  • Failure categories
  • A revision log
  • Clear safety limits

Future learning value

Why it belongs in the larger program.

  • Prompt constraints
  • Consistency
  • Model limitations
  • Evaluation
  • Authorship
  • Responsible use

Evaluation preview

Questions that keep the tool honest.

  • Does the character invent facts?
  • Does it follow conflicting instructions?
  • Can it admit uncertainty?
  • Are tests free of personal information?

Risks to inspect

What could weaken the project.

  • Treating fluent text as evidence.
  • Using personal details in prompts.
  • Changing the prompt without recording why.
  • Calling inconsistent outputs personality.

Future challenge depth

Targets, not current class instructions.

LEVEL 1

Builder

Future: define and test three character rules.

LEVEL 2

Inventor

Future: classify failures and revise one prompt constraint.

LEVEL 3

Boss Level

Future: design a safety and verification protocol for another group.