AI advised MP to reply 'congratulations' to man planning assisted dying

Australian MP Andrew Hastie says AI suggested 'congratulations' reply to man planning assisted dying suicide, highlighting risks of US-run AI systems.

Sep 18, 2026

7 cards · 1m 10s · tap to begin

From · · 1m 10s

AI advised MP to reply 'congratulations' to man planning assisted dying

Australian MP Andrew Hastie says AI suggested 'congratulations' reply to man planning assisted dying suicide, highlighting risks of US-run AI systems.

In brief

Australian MP Andrew Hastie says AI suggested 'congratulations' reply to man planning assisted dying suicide, highlighting risks of US-run AI systems. Australia needs local AI control to prevent security risks from foreign tech. Originally reported by the Guardian.

AI suggested 'congratulations'

Andrew Hastie shared Microsoft Copilot advising him to reply 'congratulations' to a man planning assisted dying suicide.

“Microsoft Copilot suggested he reply congratulations, great to hear from you, or that is wonderful news”

Australian AI push underway

Australia wants to bring OpenAI and Anthropic to local AI development to reduce security risks.

Cyber security concerns

Cyber security concerns

Australian Signals Directorate chief says cutting US AI could leave Australia vulnerable to cyber attacks.

“If a digital adversary has access to newer models, they may exploit vulnerabilities before we identify them”

Local data centers

Government plans to build Australian data centers disconnected from US to keep AI secure.

“We do not want ai tools located offshore we want compute, application, data, and decision making based in australia”

Copyright challenges

AI companies want access to all Australian content but face copyright reform hurdles.

“As australias copyright framework protects worldwide material, companies would need to strike deals with billions of rights holders”

Urgent timeline

Urgent timeline

Australian officials say copyright reform and AI rules must be finalized by end of 2026 to avoid losing global AI access.

“Decisions around training are urgent, providers will look elsewhere if we delay”

Security risks

US-based AI could be exploited by malicious actors before australia can detect vulnerabilities.

“If a digital adversary has access to newer models, they may exploit vulnerabilities before we identify them”

What matters most

Australia needs local AI control to prevent security risks from foreign tech.

Read the original on the Guardian

React

Sign in to react and comment.

Comments (0)

Life is short. Keep it sweet. Respect others' opinions and be kind!

    Recommended next

    More decks on andrew hastie and related topics.