The AI model “too dangerous” for the public

The AI model “too dangerous” for the public

From Download This Show by ABC Australia

April 17, 2026 · 29 min

About this episode

The episode discusses the dangers of the AI model Claude Mythos, the security of Signal, and the implications of blocking the Wayback Machine.

Anthropic claims its latest AI model, Claude Mythos, is “too dangerous” to release given its exceptional cyber hacking capabilities . Is it all a PR stunt, or is this a genuine threat? Plus, the FBI extracted messages from the secure messaging app Signal after exploiting a loophole involving iPhone notification databases. So how secure is Signal? And how might the loophole be closed? Meanwhile, several major global news sites are blocking the Wayback Machine , an initiative of the Internet Archive. What threat does it pose to these sites? And what will it cost us to lose this information preservation? GUESTS: Justin Hendry -- technology and business reporter and editor of Innovation Aus. Fergus Halliday -- technology & gaming journalist at reviews.org Australia ------------ This episode was produced on the lands of the Burramattagal people. ------------ Get in touch: We'd love to hear from you! Email us at downloadthisshow@abc.net.au Find all the episodes of Download This Show on the ABC Listen App or wherever you get your podcasts.

People in this episode

Guests: Justin Hendry, Fergus Halliday

Topics covered

  • AI models
  • cybersecurity
  • secure messaging
  • information preservation
  • media censorship

Keywords

  • Claude Mythos
  • cyber hacking
  • Signal security
  • FBI loophole
  • Wayback Machine
  • information loss
  • media censorship

Mentioned in this episode

Organizations: Anthropic, FBI, Signal, Wayback Machine, Internet Archive

Places: Burramattagal

More episodes of Download This Show

Explore listener stats, chart rankings, contacts and more on the Download This Show podcast page.