Lesson 3: Trust and AI
False Information
We have seen examples of where AI can be wrong, which we call hallucinations.
Hallucinations are unintentional. They are errors that we do not want. However, people can intentionally (on purpose) use and share AI responses and media that contain false information.
If one of your classmates showed you this picture below and told you there was a mountain goat on the Sydney Opera House, you might be confused or surprised. You might want to share this news with your family or friends.
But is this true or false information? And should you share it as if it was true?
The example above is silly and harmless. But sometimes the sharing of misinformation with AI can cause harm.
How can AI spread misinformation?
There are two types of false information: misinformation and disinformation.