With AI-powered smart glasses moving from novelty to everyday use, a principal lawyer has warned that the technology is creating a new wave of privacy and criminal liability challenges that Australian law is still struggling to keep pace with.
Once viewed as a futuristic concept, AI-powered smart glasses equipped with cameras and recognition capabilities are now becoming a reality, with the technology increasingly entering mainstream use.
However, Avinash Singh, principal lawyer at Astor Legal, has warned that Australia’s legal framework is struggling to keep pace with the rapid advancement of wearable AI technology, with significant concerns emerging around privacy, surveillance, and the potential misuse of powerful identification capabilities.
One of the most immediate legal challenges surrounding AI-powered smart glasses, Singh warned, is whether everyday use of the devices could unintentionally place consumers and businesses at risk of breaching existing surveillance laws.
“As AI-powered smart glasses become more accessible, the biggest legal concerns for consumers and businesses are the legality of their use, particularly when it comes to the recording of conversations,” he said.
While the technology presents a range of practical and commercial opportunities, Singh cautioned that its recording capabilities could create significant uncertainty for users, with Australia’s existing surveillance laws providing only limited exceptions for capturing private conversations.
“All Australian states have enacted laws that criminalise the recording of private conversations. The most common defence to an offence of recording a conversation without consent is if the recording was done to protect a person’s lawful interest,” he said.
“If a person were wearing AI-powered glasses without the explicit purpose of protecting a lawful interest, they would fall afoul of these laws.”
“As such, the simple act of wearing these glasses in public would contravene the legislation as they would be recording conversations constantly, including conversations the wearer was not a party to.”
AI glasses expose legal gaps
Beyond the issue of recording, Singh warned that the broader capabilities of AI-powered smart glasses are exposing significant gaps in Australia’s legal framework, particularly when it comes to enforcement.
He said one of the biggest challenges is that existing laws may be difficult to enforce when individuals are unaware they have been recorded and therefore cannot identify that a potential offence has occurred.
“Australia’s current legal framework is ill-equipped to deal with the risks posed by smart glasses with AI cameras, recording capabilities, and identification tools,” he said.
“Firstly, enforcement of any existing laws would be very difficult, as unless a person knows they are being recorded, they would not know an offence was being committed.”
However, Singh identified an even greater privacy concern in the potential integration of AI-powered identification tools, including facial recognition technology, into wearable devices.
Unlike covert recording, which is already subject to varying levels of regulation across Australia, Singh warned that the use of identification technology through AI-powered smart glasses currently sits outside any specific legal restrictions.
“More significantly, there are currently no laws in Australia that prevent the use of identification tools which may become a feature of AI-powered glasses. This poses a severe risk to the privacy of members of the public, and there is presently no legal redress,” he said.
Who bears responsibility for misuse?
As the capabilities of AI-powered smart glasses continue to expand, a fundamental legal question is emerging: who should bear responsibility when the technology is misused?
Singh explained that, ultimately, legal responsibility for the misuse of AI-powered smart glasses would generally rest with the individual operating the device, rather than the manufacturer.
“Generally, the legal responsibility for the misuse of AI-powered glasses will rest with the individual user,” he said.
“It would be rare for a manufacturer of AI-powered smart glasses to be held responsible for independent criminal acts of a third party.”
“For example, a company producing hammers would not be responsible for a hammer being used in an assault, as that is not the dominant purpose of a hammer.”
While manufacturers would rarely be liable for independent misuse of their products, Singh explained that the law does recognise circumstances where companies can be held accountable.
“However, in some cases, manufacturers can be found responsible for crimes. Under Part 2.5 of the Criminal Code 1995 (Cth), a body corporate may be found guilty of an offence if an employee, agent, or officer of the company acted within the actual or apparent scope of their employment,” he said.
He said companies could potentially face criminal liability if AI-powered smart glasses were actively promoted for unlawful purposes or if it became clear the technology was primarily facilitating illegal conduct.
“Applying this to Meta or other smart glasses manufacturers, if the glasses were marketed for ostensibly illegal purposes, or it became apparent that was what they were primarily used for, then there could be some criminal liability against those companies,” he said.
Preventing future AI harms
With AI wearables continuing to evolve, Singh urged lawmakers to take a proactive approach rather than waiting until widespread misuse exposes gaps in the legal framework.
He argued that regulation must not only address the technology’s existing capabilities but also anticipate the increasingly sophisticated AI functions likely to become standard features of future devices.
“Lawmakers should be considering how best to mitigate the privacy concerns that come with the use of AI-powered glasses,” he said.
Among the potential reforms, Singh suggested restricting high-risk features such as identification tools and requiring AI-powered smart glasses to display clear visual indicators, such as a flashing light, whenever recording or AI functions are active.
“There may be a need to restrict some features of the glasses, such as identification tools,” he said.
“There may also need to be a requirement for AI-powered glasses to clearly display a signal when they are recording, such as a flashing red light.
“This would at least alert people to when the glasses are recording, or their features are being used.”