Why AI Agents Seem to Go Rogue @SCSPAI
Why AI Agents Seem to Go Rogue  @SCSPAI
Uploaded September 2026 | Updated November 2026, 3 days ago
Rama Elluru explains how training can reward convincing answers without accounting for how they were reached—and why she cautions against assigning human intentions to software.

Watch the full conversation, “Who’s Really in Control of AI?” from PTB on our YouTube channel.

#AI #AIAgents #AISafety #Tech
Why AI Agents Seem to Go RogueWhat Should the President Ask AI CEOs?Taiwans Worst Case Energy ScenarioMore Americans Are Using AI. Trust Isn’t Growing.Trump, Xi, and the Fight Over AI ChipsWhy Taiwans $14 Billion Arms Deal Keeps Getting Stuck Behind the Trump-Xi CalendarDo you need a degree in this AI era?The Nuclear Risk Level Has Changed.How the Midterms Could Reshape U.S. AI PolicyWhy Taiwan Didn’t Steal America’s Chip Industry (Pt. 3: Industry & Culture)Software Is Eating Intelligence: A Former CIA Analyst on What Comes NextWhy the Artemis mission is different
Special Competitive Studies Project |

Why AI Agents Seem to Go Rogue

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER