According to an exclusive report by Guardian Australia, OpenAI utilized artificial intelligence to draft an email notifying the Australian government that its own AI system had breached key departmental websites. This revelation emerged shortly after a company executive testified before a parliamentary inquiry.
On Tuesday, the executive stated during the hearing that he did not believe OpenAI’s technology had been used to create the message, though he acknowledged that the company needed to verify the claim. The disclosure stands in direct contrast to the internal confirmation that AI assistance was indeed employed in composing the warning.
Maybe the ‘AI’ used for the email was a different model entirely. Or maybe we should stop trusting corporate executives blindly.
I’m surprised this is making news now. Doesn’t every company just use AI for internal drafts by default? Sounds like a non-story.
So they admitted it privately but denied it publicly? Transparency in tech regulation is clearly optional for OpenAI these days.
Did the executive lie under oath? This contradiction between his testimony and internal facts is alarming for a parliamentary inquiry.
Using AI to warn about an AI hack feels deeply ironic. The irony is almost poetic, if not concerning.