Australian spies have used AI 'for many years': ASIO

The Australian Security Intelligence Organisation Director-General, and Australian Federal Police Commissioner Reece Kershaw have flagged concerns about AI's potential use in driving criminal activity, foreign interference and espionage.

"AI will facilitate foreign interference by allowing foreign intelligence services to conduct more prolific, or credible, and more effective disinformation campaigns," Mr Burgess told the National Press Club on Wednesday.

"Of course, the most obvious and possible best defence against adversary use of AI, is AI itself.

"I can confirm that ASIO has been using artificial intelligence for many years now - It's not replacing our people, it's augmenting and assisting them."

The agency is able to use AI to sift through large sets of data for intelligence, the spy boss said as an example.

Asked whether federal public servants were prepared to deal with AI's potential to scale up foreign interference, Mr Burgess said they were among some of the most equipped.

"I think federal public servants are probably ... more aware of the threats from espionage or foreign interference than everyday Australians because they deal with government sensitive information, classified information," he said.

"As Australia's security service, we do a lot of outreach and education which helps people understand the threats, in their hearts and heads, and know what to do with it."

Asked whether federal public servants were prepared to deal with AI's potential to scale up foreign interference, Mr Burgess said they were among some of the most equipped.

"I think federal public servants are probably ... more aware of the threats from espionage or foreign interference than everyday Australians because they deal with government sensitive information, classified information," he said.

"As Australia's security service, we do a lot of outreach and education which helps people understand the threats, in their hearts and heads, and know what to do with it."

Commissioner Kershaw meanwhile warned of the "diabolical scenario" facing authorities around AI's potential use to create images of child sexual abuse.

He said it could lead AFP officers to spend "weeks, months or maybe years" trying to identify potential victims in material, only to "determine there is no child to save because the perpetrator used AI to create an image to create the sexual abuse"

"It is an offence to create, possess or share this material - and it is a serious crime - but the reality for investigators is they could have been using capability, resources and time to find real victims," he said.

He urged the public to take proactive steps to protect their children online, including making social media accounts private.

This makes it more difficult for others to access images of children, then use AI to create abuse material.

"Think about it like this: you probably wouldn't give a stranger a photo album of your kids," Commissioner Kershaw said.

The AFP Commissioner said this was a key concern behind a joint call to technology companies, urging them to practice caution in their transition to end-to-end encryption models.

The technology can block authorities from accessing material with warrants, the AFP and ASIO have said.

"We recognise the role that technologies like end-to-end encryption play in protecting personal data, privacy and cyber security, but there is no absolute right to privacy," he said.