Not in the abstract. Not as a thought experiment about timelines or whether GPT-N is “really” intelligent. What should researchers, governments, and institutions actually be doing right now, given that AI systems are getting more capable faster than most governance frameworks can keep up with?
On April 28, Tzu Kit Chan is giving a talk on exactly this for our first AI Safety Hong Kong speaker event. Tzu Kit is Chief of Staff at Atlas Computing. He’s spent several years in the AGI preparedness space: ran operations for MATS, cofounded Caltech AI Alignment, served as executive director of Stanford AI Alignment, and advises AI safety groups at 45+ universities and national organizations. His academic background is in philosophy, comparative literature, and cognitive science, so his take on preparedness isn’t purely technical. He’s also been advising decision-makers in Southeast Asia on AGI preparedness, bridging the technical alignment community and the people in government and industry who are going to have to make actual policy calls about this stuff.
The talk covers the alignment problem, what accountability means when we’re talking about frontier models, governance gaps around increasingly autonomous systems, and how communities outside San Francisco can do more than spectate.
That last part is why we started this. Most of the serious AGI preparedness work is concentrated in a few cities, but the consequences won’t be. Hong Kong is a financial and regulatory hub sitting between Silicon Valley and Beijing, and the people here in tech, finance, law, and policy are going to feel the effects of however AI governance shakes out. We’d rather they helped shape it.
Tzu Kit’s talk is the first event in an ongoing series. More to come.
Virtual, April 28, 12:00–1:30 PM HKT. Open to anyone regardless of location.
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.