Generative AI and Ethics

Generative AI creates new text, images, sound, video and code from prompts. It is powerful but can be wrong or unfair, so it must be used responsibly.

Key ideas

Generative AI creates new content; conventional AI classifies or predicts, such as sorting spam or forecasting temperature.
A prompt is the instruction given to a generative tool; clearer prompts give more useful results.
Generative AI can hallucinate: it may give confident answers that are false, so check facts in a trusted source.
It may repeat biases in its training data and may not know about recent events.
Deepfakes are realistic fake audio, images or video of a person; a watermark can show content was made by AI.
Use it honestly: say when AI helped, protect privacy and keep people accountable for what is published.

Remember

Kinds of generative AI	text, image, audio, video and code generation
Hallucination	a confident but false or made-up output
Transparency	being open that AI was used and how it decides
Accountability	people stay responsible for what AI produces
Share in training data	group count ÷ total × 100%: a small share can lead to biased output

Watch out

Calling every AI tool generative: a spam filter only classifies.
Trusting a fluent answer without checking it.
Handing in AI-made work as your own without saying so.

Worked example

Is each task for generative or conventional AI? (a) predicting tomorrow’s temperature (b) composing a new tune

Working: generative AI creates new content; conventional AI classifies or predicts

Answer: (a) conventional; (b) generative