Australia’s AI Safety Institute doesn’t have resourcing ‘to be taken seriously’ internationally, experts warn

Australia’s AI Safety Institute doesn’t have the level of resourcing “to be taken seriously” internationally, artificial intelligence experts have told a parliamentary inquiry.

AI Researchers Janet Egan and Dr Sam Winter-Levy are among 16 witnesses to appear before the Joint Select Committee on Artificial Intelligence on Friday as it continues to investigate risks and benefits of an AI roll out Down Under.

Both told the inquiry that by helping frontier companies build out their commute in Australia would be a “bargaining chip” for “influence and access”.

However, Ms Egan — a former staffer in the Prime Minister’s office who now works as an AI researcher in Washington D.C —warned Australia needed to step up its government capacity.

“I don’t think Australia’s AI Safety Institute is currently resourced enough to be taken seriously on the international stage,” she said.

“It also has an incredibly broad remit, providing AI advice to the entire government.

“The $7.5 million figure is nothing compared to what other countries are throwing at this.

“Other countries are talking about ‘well, we probably need a lot more investment in government capacity’.

“It gives you the situational awareness to understand what’s happening, it gives you the ability to work on these unsolved critical questions around alignment and control.

It comes after Prime Minister Anthony Albanese last month revealed that a rogue OpenAI agent autonomously “hacked” into the Services Australia website to access Medicare statistics.

Ms Egan said while the future of AI is “highly uncertain”, Australia should expand its ability to protect citizens as the technology develops.

She said it would benefit in terms of access to frontier models but also give it flexibility and resources to deal with workforce disruptions which seem inevitable.

“I think if Australia builds out large-scale AI infrastructure, they have this resource into the future that gives them some influence over and leverage over these companies,” she said, appearing in a personal capacity.

“Ensuring we have access to the most frontier AI models, we will have a much greater chance of being able to negotiate if AI companies are reliant on our jurisdiction to develop these capabilities.