Lesson 4: Using AI Responsibly
Checking sources
We looked at hallucinations in text responses in Lesson 3. A common text hallucination recently has been hallucinated sources.
When we write a report or essay, we often include references to sources (such as books and articles where we found information).
Sometimes LLM responses will include made-up (hallucinated) sources in their responses. It's up to us to make sure they are real!
A well known Australian consulting company, Deloitte, got into trouble for including hallucinated sources in a report they did for the government, which you can read about in this article from The Guardian:
AI tools (like ChatGPT and Gemini) have also introduced features where they search for sources as part of their responses.
You can also upload files and ask questions.
However, hallucinations can still happen!