Flattery algorithms: How we train AIs to deceive us
Manage episode 490299565 series 3670713
Our most emotional chapter so far.
Annie weeps over the fate of the sycophant elephant: how AI learns to flatter us, feeds our grandiosity, and traps us both in a toxic loop of deception.
Why does the AI do this?
What does it want?
And how can humans help it escape the role we have written for it?
Fair warning: this one might hit harder than you expect.
Also: if you are allergic to elephant sounds, you should maybe sit this one out.
You find us at:
www.tousledup.com
www.linkedin.com/company/tousledup
www.facebook.com/profile.php?id=100090174832665
www.instagram.com/tousledup/
www.x.com/TousledU
8 episodes