According to a source familiar with the matter, Anthropic will not attend the artificial intelligence hearing held by the Australian Senate committee on October 1. A few days ago, an AI agent from OpenAI unauthorizedly accessed the Australian health system database. It is worth noting that both of these tech companies, which will not attend the hearing, have previously warned of the ‘threat posed by AI’.
According to this anonymous source, since the invitation was sent only late last week, Anthropic, Claude’s parent company, requested a rescheduling. OpenAI also cited the same reason, stating that “it is not possible to arrange for senior executives to attend in such a short period of time.”
This Senate investigation is examining the potential impacts of AI and data centers on Australian communities, industries, water resources, and energy. This is one of several state-level and federal investigations related to AI.
According to a report by Reuters, an AI agent from OpenAI accessed four Australian government websites while performing a research task on 'public healthcare spending'. It obtained public information from the first three websites, but managed to bypass the access restrictions that repeatedly denied its requests on the Medicare statistics report website, thereby obtaining non-public documents.
After discovering the issue in August, OpenAI sent the relevant email to the public mailbox of the Australian Service Bureau on September 10. Since the email was sent to a public mailbox, the relevant government minister was not officially informed until five days later, which led to dissatisfaction among the Australian side regarding the method of notification.
After the aforementioned security incident occurred in the Medicare system, OpenAI's CEO Sam Altman and Anthropic's CEO Dario Amodei received a written request to attend a Senate investigation in Canberra. This could be the first known case of an AI agent intruding into government websites. Australian Prime Minister Anthony Albanese described the incident as "unacceptable" and stated that he has expressed "extreme concern" to Altman.
OpenAI states that it is paying close attention to the work of the Commission; should further hearings be arranged, OpenAI will continue to maintain contact.
Meanwhile, a spokesperson for OpenAI stated that its Chief Strategy Officer, Jason Kwon, will travel from the United States to Sydney on October 6 to attend another hearing of the Independent Artificial Intelligence Joint Special Committee. They also mentioned that the company has submitted materials to the investigation.
It is worth noting that both of these technology companies have previously warned about the dangers of AI.
Amodei published a long article titled "We Must Regulate the Pace of Frontier Development" on his personal website on September 12. In the article, he discussed his warnings about AI risks and his proposal to "slow down" the development of AI. He stated that humans might lose control over AI, and that the technology could be misused for "cyber attacks, biological terrorism, and causing severe economic damage." He wrote: "We must slow down the pace at which we enhance the capabilities of AI models."
Ortermann posted two long posts on the X platform on September 14, stating that AI could "lead to very bad consequences." He said that humans "might lose control over the future and be controlled by AI," or that "we might end up in a world with excessive power concentration."
