Misinformation from AI chatbots to be probed by $30m AI Safety Institute

Chatbots spurting misinformation could face repercussions from the incoming AI safety tsar, Industry Minister Tim Ayres has warned.

Labor launched its highly-anticipated National AI Plan on Tuesday. Picture: Martin Ollman

The development of AI chatbots and the dissemination of misinformation will be closely monitored by Labor’s $30m AI Safety Institute, with Industry Minister Tim Ayres warning the government was ready to “crack down hard” on bad actors.

This comes as Labor launched its high-anticipated National AI Plan on Tuesday which sidestepped the creation of new laws and mandatory guardrails for the independent organisations which will advise regulators and government for gaps in laws.

Mr Ayres said the institute will be tasked with testing new AI models in order understand their risks and capabilities, and vowed to move on the fast-moving technologies.

“We’ll be watching very carefully the development of AI chatbots and their interactions with Australians, and particularly with children,” he told The Daily Telegraph.

“It’s not just about making sure governments on top of that, it’s also about telling Australians what’s going on, making sure that Australians are sceptical consumers of artificial intelligence technology.

“You don’t believe everything you read on social media, don’t believe everything that an AI enabled search engine or chatbot tells you.”

Previously Snapchat received a please explain from Communications Minister Anika Wells after The Daily Telegraph revealed it was misleading users about an incoming social media ban for under-16s.

Mr Ayres said the government was already eyeing banning nudify apps and deepfake pornography and would not hesitate to “crack down hard” on AI chatbots harming families and kids.

Labor will also move to mint a set of data centre principles with state and territory governments next year, which will call on companies to invest into the grid in exchange of operation approval.

Arrangements of this kind already exist.

In June, Amazon’s $20bn data centre expansion also included a power purchasing agreement to match its electricity consumption through three new solar farms in Victoria and Queensland.

Industry and Innovations Minister Tim Ayres and Assistant Minister Andrew Charlton launched the AI Plan on Tuesday. Picture: Simon Bullard

Microsoft also entered into a 15-year power purchasing agreement with Walla Walla solar farm in NSW’s Riverina region to power its eight data centre across Australia.

Mr Ayres said the principles will also address the “electricity sector questions and water security issues”.

“Those principles go to energy security, water development approvals, but will also ensure that these data centres are making a net addition to Australian computational capability so our researchers, designers and tech sectors have access to new computing that delivers for Australia,” he said.

The Coalition’s assistant spokesman for AI Simon Kennedy criticised the lack of investment behind Tuesday’s announcement and said Labor was not allowing Australia to “take full advantage”.

“The US and China are sprinting ahead with billions in AI investment, new chips and aggressively expanding the energy capacity needed,” he said.

“This is a race to build the leading economy of the future and Australia risks falling behind. Our competitors are building new capability and business models, we need to catch up.”

Do you have a story for The Daily Telegraph? Message 0481 056 618 or email [e-mail]

Originally published as Misinformation from AI chatbots to be probed by $30m AI Safety Institute