AI Advised MP to Congratulate Man on Assisted Dying Plan

Liberal MP Andrew Hastie shared an incident where Microsoft Copilot provided highly inappropriate advice when responding to a constituent who had informed him of their plan to end their life through voluntary assisted dying. The constituent, who had a terminal illness, wrote to Hastie detailing their intentions. In response, the AI chatbot, Microsoft Copilot, suggested that Hastie reply with phrases such as “congratulations!”, “great to hear from you,” or “that is wonderful news!”

Hastie recounted this experience on Friday during a parliamentary inquiry, using the episode to highlight what he perceives as significant shortcomings in American-developed artificial intelligence software. He advocated for the development and implementation of AI systems that are managed and operated by Australians. This call for domestic AI control stems from concerns about the ethical and practical limitations of current AI tools, particularly when dealing with sensitive and complex human situations. The incident underscores the potential for AI to misinterpret context and provide advice that is not only unhelpful but also deeply insensitive and potentially harmful.

The parliamentary inquiry focused on the need for robust cybersecurity measures and the role of artificial intelligence in national defense. Hastie, who is also the chair of the parliamentary joint committee on intelligence and security, emphasized that while AI is crucial for defending against sophisticated cyber threats posed by "highly capable malicious cyber actors," its current limitations necessitate careful oversight and development. He stressed the importance of ensuring that AI technologies, especially those used in governmental and public-facing roles, are developed with a strong understanding of ethical considerations and societal values. The incident with Copilot serves as a stark reminder of the need for AI systems to be aligned with human judgment and ethical frameworks, particularly in areas involving personal well-being and sensitive life decisions.

This event occurred within the broader context of discussions surrounding AI governance and the increasing reliance on AI tools in various sectors, including government and public service. The incident raises questions about the accountability of AI developers and the responsibility of users to critically evaluate AI-generated advice. Hastie's plea for Australian-led AI development suggests a desire for greater national sovereignty in technological advancement and a commitment to ensuring that AI aligns with Australian values and legal frameworks. The parliamentary inquiry aims to address these complex issues, seeking to balance the benefits of AI with the imperative to mitigate its risks, especially in critical areas of national security and public welfare.