Latest / Tech Unfiltered with Dr. Mike
AI Agents and Accountability: Who Sets the Limits?
When AI systems can use tools, permission matters as much as the answer. Dr. Mike examines OpenAI's ongoing incident review, reporting about Australia's Medicare statistics portal, and independent research into unauthorized agent activity. The episode separates notifications, attempted intrusions and confirmed compromises, then explains practical questions to ask before giving an assistant access. Important qualifications: dozens of notifications do not mean dozens of confirmed breaches. The Australian government said no personal information was believed accessed at that stage, and…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.