Intro to AI Safety @ UWA campus

Intro to AI Safety: Why Worry About Chatbots?
Monday 28 September · 12pm · UWA (venue TBA)
Free entry · Free pizza + zero-sugar soft drinks
AI can now write code, reason through problems, and increasingly take actions through computers. This year, one model found several hundred previously unknown security vulnerabilities in code that had already been tested about as hard as code gets.
So when does an unreliable chatbot become a technology worth taking seriously?
This is a beginner-friendly 30-minute introduction to AI safety, followed by open Q&A and pizza. It covers:
what people actually mean by "AI safety"
the difference between a chatbot and an autonomous agent
misuse, alignment, and societal risk — and why they're different problems
three open research problems: interpretability, evaluations, and control
the strongest arguments for and against worrying about advanced AI
the remote programs that are the actual entry routes into the field, and the people in Australia doing this work now
No prior AI safety knowledge required. The main talk is accessible to anyone interested in AI; there's technical depth in Q&A for people who want it.
Sceptics especially welcome. This isn't a pitch for one prediction about the future. It's an invitation to look at the arguments and evidence together — and the hostile questions are the interesting ones.
Stay afterwards for pizza and to meet other people around Perth interested in AI, ML and safety.
The one thing to do afterwards: join the Discord.
Run by a new AI safety group at UWA, operating under EA UWA.

LOCATION
CCZ UWA
instruction video:
https://www.instagram.com/reel/DHaDMJKSH0Z