Independent Member for Curtin Kate Chaney MP is calling on the Government to urgently and substantially increase funding for the Australian AI Safety Institute (AISI), after a week in which the leaders of the world's most powerful AI companies admitted that the technology is developing faster than we can control it.
"Anthropic CEO Dario Amodei is urging the industry to slow itself down because AI capabilities are now improving so fast that they risk outrunning our ability to understand and control these systems," Ms Chaney said.
"When you have Elon Musk and the heads of Anthropic, OpenAI and Google DeepMind agreeing publicly that the risks posed by AI are careening out of control, the whole world needs to pay attention. And here at home, we should be asking: how can we preparing ourselves for this risk? Who is protecting the interest of Australia and Australians?
"The answer, right now, is barely anyone. The Albanese Government seems happy to leave the Australian community and economy exposed to the risks of this emerging technology – from cyber-security, bioweapons, supercharged disinformation campaigns, and AI agents escaping human control. Donald Trump and tech billionaires cannot be relied upon to make sure AI models and agents are well-tested and safe for use. The Government must act.
“Australia's AI Safety Institute is our frontline defence against exactly the kind of catastrophic risk the industry is now sounding the alarm on. Its responsibility is to test new AI models and agents for risks, work with the Government to develop appropriate responses and collaborate with international AI safety institutes. Yet, AISI is allocated less than $8 million a year. This is woefully inadequate.
"Everyone I have talked to across the AI industry, business, civil society, experts and community, agrees that the AISI needs a big injection of resources to keep up with the incredible pace of AI developments. The UK's equivalent institute gets sixteen times more funding, and it shows - it is widely recognised as the international gold standard for AI safety.
"I'm calling on the Government to increase AISI funding to $100 million a year. This is a modest sum against the Federal Budget. It is small change next to the value these AI companies are extracting from the Australian market and the damage they could cause if potential risks come to pass.
“This much-needed funding boost would build the AISI’s technical capability to properly test frontier models for catastrophic risks, strengthen the institute’s position to recruit AI safety experts, and set up the opportunity for Australia to pull its weight internationally in AI testing.
"I raised these issues in my AI Discussion Paper in May, where I also made the case for giving AISI a legislated power to compel pre-deployment testing, rather than relying on voluntary cooperation from companies.
"Some AI capabilities are now doubling roughly every four months. An underfunded institute with limited testing capacity, operating without robust powers, is not a serious answer to an accelerating risk the industry itself has identified. We are not talking about hypotheticals anymore.
"Australia’s protection from frontier AI risks cannot rely on overseas companies and foreign presidents. We don't get to opt out of the consequences if something goes wrong. Properly resourcing the AI Safety Institute is the minimum the Government should be doing right now. After the revelations of this week, this becomes even more urgent."
[ENDS]
Media enquiries: Cade Smith | 0408 997 003 | [email protected]