← Back to course
5/6
Week 09 Β· Real, Fake and Whose?

Friday

Impact Friday: Trust and credit
// Information integrity, provenance and intellectual property
⏱ about 20 min

Friday: Impact Friday: Trust and Credit

This morning, a real notice goes up on the library doors: "Library closed this afternoon for painting." Librarian Joss wrote it herself.

After lunch, students are still walking up to the locked doors. "Probably another prank," one of them says.

Comet groans. "The fake flyer made people stop trusting the real one!"

Wren nods. "What does the evidence say? One fake did damage after it was gone."

Teacher Lin joins them with QZ-5. "And I only spotted the copied paragraph because I know that reader by heart."

Nova projects risk rows R5 and R6. "Would you like a hint?" she asks. "Use what you learned this week to rate both."

Wren hands you the pencil. "Our reviewer rates first."

When a fake spreads

NIST says false content, including content made with generative AI, can make people trust true evidence and information less.

The painting notice was real. The fake flyer made people doubt it.

This is a made-up example from the SH-1 story, but the pattern is the one NIST describes.

TRUST AND FAKES
  • Read the question.
  • Tap your answer.
Why did students ignore the real painting notice?
How could the "Made with SH-1" label help trust?

Laws can help and harm

The CSTA standards point out that laws govern privacy, data, property, information and identity in computing.

These laws can help and harm. They can speed up or slow down advances, and protect or go against people's rights.

This course does not name or explain any law. The crew's own habit is simpler: give credit, and ask permission.

So the panel will ask each teacher before their notes are uploaded to SH-1.

StatementTrue or false?
Laws about computing can have both helpful and harmful effects.?
Laws can protect people's rights, and they can also go against them.?
Asking a teacher before uploading their notes is part of the crew's practice.?
Giving credit only matters for students, not for AI systems.?
WHY THIS EXERCISECredit and permission are habits the crew wants SH-1 to follow too.
AI actor spotlight: domain experts
NIST says domain experts know the field where an AI system is used.
They help design the system and help make sense of its outputs.
At Harbor Point, the teachers are the domain experts. Teacher Lin spotted QZ-5 because she knows the Biology reader well.

Rate R5 and R6

Risk combines how likely a harm is with how big it would be.

R5, information integrity. Likelihood: Medium. SH-1 confabulates, and unlabelled outputs could be shared as real. Harm: Medium. People could act on wrong facts, but labels and teachers can catch them.

R6, intellectual property. Likelihood: Medium. QZ-5 shows SH-1 can repeat a source. Harm: Medium. Copying without credit is unfair to authors, but teachers checking quizzes can catch it.

RowRiskSH-1 exampleLikelihoodHarm
R5Information integrityUnlabelled SH-1 text and images shared as realMediumMedium
R6Intellectual propertyQuiz copies a textbook paragraphMediumMedium

Week review

  1. High-integrity information separates fact from fiction, admits uncertainty, says how it was checked and can be traced.
  2. Misinformation is spread by mistake; disinformation is made to deceive. Tiny changes can fool people.
  3. Deepfakes are highly realistic synthetic media. Fakes can make people trust real notices less.
  4. Provenance records where content came from, with visible or hidden marks and metadata.
  5. Generative AI can make copying and plagiarism easier. How copyright fits with AI is still debated.
  6. R5 and R6: likelihood Medium, harm Medium.
WEEK CHECK
  • Read the question.
  • Tap your answer.
The crew rated R6, intellectual property. What were the ratings?
Which label did the crew propose for every SH-1 output?
On paper, add rows R5 and R6 to your risk register, with each rating and one piece of evidence.

Outstanding week, reviewer. Tomorrow is the Explorer Quest: a family provenance hunt.

← Thursday