The federal police union has said a single portal should be set up for victims to report AI deepfakes to police, lamenting that police have been forced to “improvise” laws to charge the first person to face prosecution for spreading deepfake images of women last year.
Attorney General Mark Dreyfus introduced legislation in Parliament in June that would create a new criminal offence for sharing, without consent, sexually explicit images that have been digitally created using artificial intelligence or other forms of technology.
The Australian Police Federation Association (AFPA) supports the bill, arguing in a submission to a parliamentary inquiry that the current law is too difficult for officers to use.
They pointed out the case of a man who He was arrested and charged in October last year for allegedly sending deepfake images to Brisbane schools and sporting associations. The eSafety Commissioner separately launched proceedings against the man for failing to remove “intimate images” of several prominent Australians last year from a deepfake pornography website.
The man was fined $15,000 as part of the civil contempt of court case. His criminal and civil cases are ongoing, and the civil case will return to court in August.
“Due to limited resources and the lack of specific and direct legislation relating to fake sexually explicit material, investigators were forced to ‘improvise’ offences to prosecute,” Afpa said. “Six further charges relating to ‘obscene publications and programmes’ were brought against (the man).”
The internet freedoms advocacy non-profit Electronic Frontiers Australia told Guardian Australia in May that parliament should not rush into granting new powers to the police until this case had determined whether current powers were adequate.
Afpa said the eSafety method of filing a civil suit also had disadvantages because it is expensive and there was a “good chance” that the infringer was a “low-income, low-asset person” who is therefore “effectively immune from civil proceedings.” If they can find out who created the image, that is.
“It is often impossible to determine who distributed the images; offenders are often very technologically savvy and are good at covering their tracks to avoid prosecution.”
Afpa said it is often difficult for police to determine who the victim is, whether they are a real person and where they are located, meaning deepfake investigations can take “countless hours.”
After the newsletter promotion
“With the rise in the creation of deepfake child exploitation material, the role of law enforcement and identifying a victim is becoming exponentially more difficult,” the union said. “How much time do investigators spend trying to find a child who may not even exist or whose image was stolen but who ultimately has not been abused?”
It is also difficult to determine where the image was first created, Afpa said, noting that people often use virtual private network (VPN) connections to hide their location.
While victims can currently report to the e-Safety Commissioner when an image, real or not, of them has been shared online without their consent, Afpa said this model should be revised to allow victims to report directly to the authorities.
AFP has suggested that the Australian Centre for Countering Child Exploitation, run by the AFP, could assess initial reports and then share them with the relevant state or territory police force for further investigation.
This would also help victims to report cases, the union added, because many find it traumatic and difficult to enter a police station with sexually explicit images to report it to the police.
Afpa argued that the legislation should be accompanied by an educational campaign to reduce the stigma around reporting and educate the public about deepfakes, in addition to the reporting portal.
The committee will hold its first hearing on the legislation next week.
