AI systems are getting more capable. Are they also getting less aligned? Join us for the second AI Safety Hong Kong Reading Group, where we’ll dig into one of the most important questions in AI safety: what happens when a system gets better at achieving goals, but not necessarily the ones humans actually want? Emerging evidence suggests that some frontier models may already be showing worrying behaviours. We’ll discuss a short, accessible article 'Misaligned AI Is No Longer Just a Theory' and explore questions regarding what misalignment means, which examples are compelling, and when unexpected behavior becomes a safety concern. No technical background required. Whether you work in policy, governance, research, law, business, or you’re simply trying to make sense of where AI is headed, you’re welcome.