← Back to lesson

📮 Homework · OpenAI agent 'infiltrated' an Australian government website

One card per section. Copy the prompt into any AI chatbot at home. Every task has an offline fallback so no AI account is needed.
Filter by key stageClick one or more chips to filter the sections below.
§1

Myth vs Reality: did OpenAI tell Australia straight away?

Y9Y10Y11Open section ↗
🏠 Take-home task

Bring in one AI security-incident headline

📋 Copy this into any AI chatbot
Please suggest three real UK news headlines from this week about AI security incidents, breaches, or safety warnings. Do not tell me which are the most or least serious — I want to work it out from what the underlying article actually says about the timeline.
Find one AI security-incident headline from any UK news source this week and bring it in. Write two sentences: one saying whether the headline is a Myth or a Reality by the standard we used today, and one saying what the article did or did not tell you about the timeline between breach and disclosure.
AI useOptional. If you use an AI you can ask it to find recent UK AI security headlines. Any of Copilot, ChatGPT, Claude, Gemini is fine.
If no AI or internetYou can use a physical newspaper or a school-provided printout.
Bring backOne printed or handwritten headline plus two sentences.
Stay safeNever share your name, school or address with the AI. If the headline covers something you find worrying, talk to a trusted adult.
§2

Reading between the lines: Albanese's 'very frank discussion'

Y9Y10Y11Open section ↗
🏠 Take-home task

One more implicit assumption

📋 Copy this into any AI chatbot
Please read the BBC News article at https://www.bbc.com/news/articles/c6vgy0333dppo and list three implicit assumptions the article makes without stating them outright — not from the 'very frank discussion' paragraph, from elsewhere in the piece. Do not give the reasons.
Read the full BBC article at home (https://www.bbc.com/news/articles/c6vgy0333dppo). Find one more implicit assumption the article makes elsewhere in the piece — outside the 'very frank discussion' paragraph. Write one sentence naming the assumption and one sentence saying how the article sets it up.
AI useOptional. If you use an AI you can ask it to list common implicit assumptions in political reporting. Any of Copilot, ChatGPT, Claude, Gemini is fine.
If no AI or internetYou can complete this with the printed article and a pen.
Bring backTwo sentences, handwritten or typed.
Stay safeNever share your name, school or address with the AI.
§3

Debate: should AI companies have a legal 72-hour disclosure deadline?

Y10Y11Open section ↗
🏠 Take-home task

The argument the class missed

📋 Copy this into any AI chatbot
Please summarise the BBC News article at https://www.bbc.com/news/articles/c6vgy0333dppo in three sentences, then name one argument on either side of the motion 'AI companies should be legally required to disclose agent security incidents within 72 hours' that is only implied and not stated outright.
Read the full BBC article at home (https://www.bbc.com/news/articles/c6vgy0333dppo). Find one argument, on either side of the motion, that did not come up in the class debate. Write two sentences: one naming the argument, one saying whether it would have changed your vote.
AI useOptional. If you use an AI you can ask it to summarise the article or suggest arguments the class might have missed. Any of Copilot, ChatGPT, Claude, Gemini is fine.
If no AI or internetYou can complete this with the printed article and a pen.
Bring backTwo sentences, handwritten or typed.
Stay safeNever share your name, school or address with the AI.