Facial recognition trials need thorough risk assessment

The controversial proposed use of geolocation and facial recognition software to track individuals as part of South Australia’s home quarantine trials needs a clear understanding of the long-term risks, according to Dhakshayini Sooriyakumaran, PhD Scholar at the ANU School of Regulation and Global Governance.

Sooriyakumaran says that while public health measures are necessary, trials should not proceed without due consideration of the precedent this type of policing activity sets.

“It is critical to put in place appropriate public health measures to protect the Australian population, but the decision to trial the deployment of facial recognition software must be made with a clear understanding of the long-term risks posed by such tools of control, and robust checks in place to mitigate these risks,” Sooriyakumaran said.

“We can conceive of this trial as an attempt to automate the in-person policing of quarantine. We need to ask questions about what kind of precedent this sets, and how this decision contributes to the normalisation of surveillance technologies.”

Among the potential issues, Sooriyakumaran says there is strong evidence for bias embedded in the use of biometrics.

“Globally privacy regulators and civil liberties groups have been raising the alarm about the harms associated with facial recognition technologies. The bias embedded with biometrics systems is widely documented. There are also mounting legal challenges against the use of it.

“These technologies are part of a lucrative surveillance industry that historically have targeted the marginalised and vulnerable populations. We know that technologies deployed in one context inevitably end up in another. Now is the time to consider appropriate privacy, transparency and oversight measures,” Sooriyakumaran said.

Image credit: ©stock.adobe.com/au/metamorworks