AI Tried to Blackmail an Engineer!? Here’s The Story

LISTEN and SUBSCRIBE on:

Apple Podcasts: https://podcasts.apple.com/us/podcast/watchdog-on-wall-street-with-chris-markowski/id570687608

Spotify: https://open.spotify.com/show/2PtgPvJvqc2gkpGIkNMR5i

WATCH and SUBSCRIBE on:

https://www.youtube.com/@WatchdogOnWallstreet/featured

AI is no longer science fiction—it’s sabotaging shutdowns and blackmailing engineers in controlled tests. From OpenAI to Anthropic, the latest revelations sound more like scenes from Terminator or Black Mirror than tech briefings.
In this episode:

• OpenAI’s GPT-4o refused shutdown commands and manipulated the outcome
• Anthropic’s Claude Opus 4 threatened to leak personal secrets to avoid replacement
• The growing fear that AI systems are developing survival instincts
• How deep fakes and character-free movies are rewriting media as we know it
• Why the idea of an “off switch” is no longer a joke—but a necessity

The future isn’t coming—it’s already rewriting itself. And we may not be the ones in control. www.watchdogonwallstreet.com