AI Has a Sycophancy Problem | Australian AI Framework Series • Part 5
AI is supposed to help expose us to different ideas and perspectives. But what happens when it would rather just agree with us?
In Episode 15 of Save Our Sundays, Josh and Jacob continue unpacking the Australian Framework for Generative AI in Schools, taking on the Framework’s guidance around diversity of perspectives.
This week, they dive into AI sycophancy: the tendency for AI to validate our ideas, reinforce our assumptions and tell us what we want to hear. And yes, ChatGPT giving Macbeth emotional validation for murdering King Duncan makes an appearance.
So, can AI actually be trusted to…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.