You bring an idea to an AI assistant. It finds something encouraging to say. The answer sounds intelligent, but did it help you think?

Before deciding how intelligent a machine is, this talk starts with what we mean by intelligence in people. Using multiple intelligences as a lens, we explore the uneven profile of AI: impressive in some situations, surprisingly limited in others. As its capabilities change, recognising those limits remains only part of the work. We also need to recognise what we are asking it to give us.

The courtier’s flattery is the trap. Intrapersonal intelligence offers a way to notice our own desire for reassurance and choose a more useful conversation. Instead of asking for approval, we can ask the assistant to find weaknesses, explore failure and argue the other side.

A session for anyone who wants an AI conversation to improve an idea, even when that means hearing something uncomfortable.

For non-technical audiences and professionals interested in how AI works and how it influences their judgement. Familiar examples connect uneven machine capabilities with the human habit of seeking confirmation. The session gives participants a concrete way to invite challenge and examine their own questions. Suitable for AI literacy, education and interdisciplinary programmes. Technical knowledge is not required.