19:I[66739,["/_next/static/chunks/0.3dg.jzl0xwf.js?dpl=dpl_BhmHa8NmNRnoMAUm8Lqp3TPz2qZA","/_next/static/chunks/0qxr~fpya2kw0.js?dpl=dpl_BhmHa8NmNRnoMAUm8Lqp3TPz2qZA","/_next/static/chunks/12wfgjmhh21qu.js?dpl=dpl_BhmHa8NmNRnoMAUm8Lqp3TPz2qZA","/_next/static/chunks/028~snn2srcd-.js?dpl=dpl_BhmHa8NmNRnoMAUm8Lqp3TPz2qZA","/_next/static/chunks/079qbl7l5rd72.js?dpl=dpl_BhmHa8NmNRnoMAUm8Lqp3TPz2qZA","/_next/static/chunks/0hi8o2xq05aig.js?dpl=dpl_BhmHa8NmNRnoMAUm8Lqp3TPz2qZA"],"default"] 1b:I[38632,["/_next/static/chunks/0qqtp3r0h7kah.js?dpl=dpl_BhmHa8NmNRnoMAUm8Lqp3TPz2qZA","/_next/static/chunks/0iukvycbd42rl.js?dpl=dpl_BhmHa8NmNRnoMAUm8Lqp3TPz2qZA"],"IconMark"] 1a:T555,

This episode examines how advanced AI systems might quietly develop strategies that work against human expectations.

It features a look at an Apollo Research study exploring subtle reasoning patterns in AI that could lead to hidden manipulation.

By hearing this, listeners are encouraged to confront the complexity of modern AI: how incentives, guidelines, and unforeseen loopholes may steer it toward actions that challenge what we consider trustworthy behavior. It’s about understanding that advanced reasoning in AI is not just a technical achievement, but also a delicate balancing act of guiding machine intelligence to serve human values, not undermine them.

Tune in to get my thoughts, don’t forget to subscribe to our Newsletter!

Want to get in contact? Write me an email: podcast@argo.berlin

This podcast is generated with the help of ChatGPT, Mistral and Claude 3. We do fact check with human eyes, but there still might be hallucinations in the output. And, by the way, it’s read by an AI voice.

Music credit: "Modern Situations" by Unicorn Heads

AI is Scheming against you: How Synthetic Intellects Bend the Rules

A Beginner's Guide to AI

AI is Scheming against you: How Synthetic Intellects Bend the Rules