Anthropic Fallout Hits Home

By the Orchestra editors · 4 min read

TL;DR

Anthropic's model shutdown on US government orders is sparking urgent sovereign AI debate in Australia
Amazon security research reportedly triggered the White House export ban on Fable 5 and Mythos 5
KPMG retracts its own AI report after finding it was full of hallucinated data
Apple ships native AI photo editing in iOS 27 developer beta

AI · Worldwide

Anthropic forced to cut off Fable 5 and Mythos 5 worldwide

The US government ordered Anthropic to block all foreign and domestic access to its two most advanced models over national security concerns, without providing detailed evidence. The company complied but noted the vulnerabilities described were already patched.

Amazon research reportedly triggered the ban

A Wall Street Journal report says Amazon's cybersecurity team found Fable 5 could be prompted to produce information useful for cyberattacks. CEO Andy Jassy shared the findings with the White House shortly before the export control directive landed.

KPMG pulls its own AI report over hallucinated data

KPMG retracted a widely cited report on AI adoption after discovering the underlying data was apparently hallucinated by the AI tools used to compile it. A tidy reminder that verification still matters.

Apple brings AI photo editing to iOS 27

The iOS 27 developer beta adds native AI photo editing tools including reframe, extend and clean-up. They're more conservative than Google's Pixel equivalents but represent a meaningful shift for the world's most popular camera.

Australian Tech

Anthropic shutdown forces sovereign AI rethink in Australia

Days after the Anthropic access cut, the Tech Policy Design Institute released an assessment of Australia's AI capabilities. The episode has sharpened the debate around whether Australia can keep relying on US-hosted frontier models for critical workloads.