Nurses warned against invasive surveillance trade-offs when adopting AI.

Law reform and collective action are needed to set boundaries on the use of digital tools to monitor workers, an expert on human rights and new technology told the NSWNMA annual conference.

Edward Santow, director of policy and governance at the Human Technology Institute, UTS, said common IT tools have hidden surveillance features that can be configured to keep tabs on workers beyond communication needs.

Workers should not have to accept harmful surveillance as an integral part of beneficial artificial intelligence, he added.

Santow, a former Australian Human Rights Commissioner, warned against “fantastic AI systems that are also monitoring or surveilling you at every moment of your workday”.

“Sometimes the argument is made that you can’t have these wonderful aspects of AI without putting up with these terrible things.

“And my argument to you is that it’s bullshit. (…) We don’t have to accept that the technology that (has) genuine, real benefits for people must also cause harm through, for example, over-surveillance of workers.”

Santow said the most common piece of technology being used to monitor workers in Australia today is Microsoft Teams.

“There are buttons that you can press behind the scenes by a manager or as an employer that can monitor you in a whole bunch of different ways.”

A conference delegate asked Santow how to find out what level of surveillance an employer is using on Microsoft Teams.

He replied, “You can ask the question of your manager, but ultimately this needs to happen at the system level. (…) That’s something that’s really important for representatives to actually take very seriously and ask those questions.”

He said a “very small tweak” in the law is needed to make clear that “monitoring should be no more than is absolutely necessary to, for example, keep people safe” – regardless of what your employment contract says.

“It makes sense to have some monitoring in dangerous environments like mines to make sure that workers there are safe, but that doesn’t justify people’s conversations being recorded and being told, ‘I can see you’ve got this particular political view, and we don’t really like that in our company.’”

Santow said about 80 per cent of AI projects fail – mainly because the people designing and implementing them don’t understand the field in which they’re being integrated.

“You may know a lot about the science and the technology, but if you don’t understand the area in which a nurse or another health professional operates, and you just want to give them a piece of tech, there’s a really good chance it’ll fail.”

He added, “Technology should make your job easier or make you better at doing your job.

“You should not have to completely change the way in which you do your job to suit the needs of the machine.”