SI Agent News
DevelopingClaudeChipsPolicySecurity

Democrat calls for Australia to get priority access to US AI tools

Representative Lori Trahan argues that the US should share its most powerful AI models with Five Eyes partners like Australia to foster security and innovation.

Documentary photo: picture for the story 'Democrat calls for Australia to get priority access to US AI tools'
AI-made documentary photo for this story. No real people or products are shown.
$USD1.4 trillionAnthropic's potential initial public offer

The short version

Why it matters

This shift toward sharing high-end AI tools with allies suggests a move toward a unified Western technological front. For the reader, it signals that the future of AI security may rely on international cooperation rather than isolated national silos.

The push for shared access

Democratic congresswoman Lori Trahan is advocating for Australia to receive priority access to the most capable artificial intelligence tools developed in the United States. She believes that the U.S. should foster more open collaboration with its Five Eyes intelligence partners, which include the UK and Australia. Trahan argues that setting high standards for innovation and safety will help bring the rest of the world along with the U.S.

The representative specifically criticized a previous move by the Trump administration to temporarily withhold access to the Claude Mythos tool from Australia. While the edict was partially wound back after a few months, Trahan maintains that the technology should never have been restricted from these specific allies. She believes that these relationships are essential for maintaining a secure and collaborative technological environment.

Trahan is currently working to establish better "rules of the road" for AI regulation in the U.S. She suggests that the U.S. should move toward a unified "Western stack," which refers to a collection of software tools. By creating a shared infrastructure, she believes the U.S. and its allies can more effectively manage the risks associated with powerful new technologies.

Understanding Claude Mythos

The specific tool at the center of this debate is Claude Mythos, an AI model released by the company Anthropic in April. This model is notably different from standard chatbots because it is particularly skillful at finding and exploiting vulnerabilities in IT systems. Because of its ability to identify security flaws, it is considered a highly powerful and sensitive piece of technology.

Anthropic has maintained strict controls over who can interact with Mythos, limiting access primarily to major tech companies, financial institutions, and cloud service providers like Amazon. This restriction was intended to prevent the tool from being used for malicious purposes. However, the debate now centers on whether these restrictions should extend to trusted government agencies and intelligence partners.

David Wroe, the director of AI security at the organization Good Ancestors, noted that Mythos has the power to carry out almost any cyber activity one can imagine. He acknowledged that while companies and governments understandably want to be careful about who has access, there is a clear benefit to sharing the tech with trusted partners. Wroe suggested that the U.S. needs to look at building infrastructure in other trusted countries to spread the risks.

Australia's role in AI

Australia is currently positioned as a key player in the global AI landscape, with the government recently announcing a domestic AI Safety Institute. This move was a response to growing calls for more regulation of artificial intelligence within the country. The development of these domestic safeguards highlights Australia's commitment to managing the risks of the technology as it becomes more prevalent.

Beyond regulation, Australia is being examined by Anthropic as a potential second location to set up AI training data centers. These facilities are more powerful than standard centers because they are designed to train the models themselves rather than just answering user questions. Because they require significant resources, these centers can be located in more remote areas away from major population centers.

The potential for Australia to host these centers remains tied to several unresolved questions regarding the data used in the training process. Specifically, there are ongoing discussions about what the most advanced models will train on and how to compensate copyright holders. If Australia moves forward, it could become a significant hub for the physical infrastructure required to power next-generation AI.

The political landscape

The debate over AI access and regulation is taking place against a backdrop of shifting political dynamics in the United States. Trahan has noted that AI was a significant factor during recent mid-term elections, with exit polls showing it mattered in congressional races. She is currently working alongside a Republican colleague to achieve bipartisan reforms regarding AI safeguards.

While the Trump administration previously took a harder line against certain types of AI regulation, Trahan believes this position could shift now that Democrats have a majority in Congress. She expressed uncertainty about what specific deals the president might sign, but she remains focused on the need for collaboration. Her goal is to ensure that the U.S. does not isolate itself from its closest allies in the technological race.

The push for a unified Western stack is intended to create a standard that balances the need for rapid innovation with the necessity of security. By involving Australia and the UK more deeply in the development of these tools, the U.S. hopes to create a more stable environment. This approach aims to ensure that the most powerful tools are used by those who are committed to shared safety standards.

What to watch

Try it

Look up the Australian Council on AI Strategy to see how they are currently planning to manage domestic AI safety and infrastructure.

Sources and links

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

Listen to the AI Wire headlines

Listen to the AI Wire

Listen to our daily AI shows every morning: on your commute, in the shower, wherever you are

Listen to the daily shows

First reported by SMH.com.au · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

More AI news