Workers could be closely monitored and potentially discriminated against without appropriate oversight of artificial intelligence in the workplace, unions say.

The Australian Council of Trade Unions has called for a national AI authority to be introduced to regulate the adoption of the technology in workplaces.

But the Business Council of Australia has cautioned against introducing more red tape, saying it could prevent businesses from having a “competitive edge” over overseas companies.

A parliamentary inquiry into the digital transformation of workplaces is examining the benefits and risks of automated decision-making in Australia.

Artificial Intelligence could boost the economy

US multinational financial services company Goldman Sachs said AI had the potential to boost global annual productivity by 1.3 per cent.

The International Monetary Fund said about 40 per cent of jobs worldwide were exposed to AI and about 60 per cent of jobs in advanced economies were likely to be impacted.

About half of the exposed jobs could benefit from AI integration and enhanced productivity, according to the IMF, but the rest could be replaced.

Business Council of Australia chief executive Bran Black told the ABC that Australia needed to embrace AI because it gave local businesses a “competitive edge”.

Mr Black said AI was already being used to make workplaces safer or more productive.

“Some of our members in financial services … have loan assessors, and those loan assessors once upon a time would have had to spend hour upon hour going through loan materials,” Mr Black said.

“A lot of that work is now being done using AI but the important part of that is that it gives the assessors the scope to focus on the value adding part of their role.

Mr Black said it was important that Australian businesses could catch up to their overseas competitors who have raced to adopt AI in their workplaces.

“So for us here in Australia, it’s important that we try and as much as we can, realise the same opportunity so that we’re not left behind,” he said.

ACTU assistant secretary Joseph Mitchell said the uptake of AI has been useful in some industries, such as healthcare.

Mr Mitchell said AI had the potential to remove the bulk of administrative work for health workers.

“For example in the health industry … [AI can] improve the interpretation of scanning, to reduce the significant administrative burden that’s placed on our health professionals,” Mr Mitchell said.

Workplace surveillance, AI discrimination a risk for employees

However, Mr Mitchell said AI had the potential to enable workplace monitoring, increase work intensity and lead to discrimination in the recruitment process.

He said health workers who saved time on administrative tasks with AI could develop burnout if they replaced that with additional patient contact hours.

“That reflection and that time taking clinical notes and documenting the patient’s journey … if that’s just replaced with additional patient contact hours, we’re really worried about leading to burnout,” he said.

“We think there’s a balance to be struck between productivity enhancing technology and burnout and stress and work intensification.”

Mr Mitchell said unions were worried that AI could also lead to workplace surveillance.

“Things like their location, how often they go to the toilet, what they’re doing with their hands at any given time, where they’re looking and who they’re talking to,” he said.

“Even things like rostering decisions where workers have no visibility over what characteristics are used to determine whether or not people are put on shift and what kind of shifts they’re given.

Businesses needed to consult with workers before introducing AI into the workplace, Mr Mitchell said.

“What typifies the positive introduction is that workers are part of the story … when they come to introducing AI,” he said.

Andreas Cebulla from the Australian Industrial Transformation Institute said it was unknown how AI integration could affect how much people were paid.

“What we don’t know is how this is going to affect how jobs are rated and literally paid in a workplace,” Dr Cebulla said.

“As the tasks change the position of someone in an employment hierarchy will change as well.”

Dr Cebulla said AI needed to be introduced into the workplace “collaboratively” so everyone understood how it would be used.

“Once you’re introducing a tool, there will be people using it so you need to teach them,” he said.

Does Australia need to regulate the use of AI at work?

In September, the Australian government proposed introducing 10 “mandatory” guardrails for high-risk AI.

The government said it would look at three regulatory approaches, from changing existing regulations to introducing a cross-economy AI specific law.

In a submission to the Inquiry into the Digital Transformation of Workplaces, the ACTU called for the establishment of a national artificial intelligence authority.

The ACTU said the authority would provide a “coordinated national and unified policy approach that balances innovation, development, risk, regulation, and rights”.

Mr Mitchell said Australia needed a centralised means to ensure AI did not impede on workers’ rights.

“We don’t want to see different arms of government working in different directions and creating gaps in the process when they’re responding to issues that are affecting them,” he said.

But Mr Black cautioned against introducing new AI regulations and said businesses could work with existing laws surrounding surveillance and discrimination.

He said businesses were already burdened with enough red tape.

“The Fair Work Act already has pretty significant provisions in place that deal with those types of issues,” he said.

“We should be relying on our existing regulation before considering the need for new regulation.”

Dr Cebulla said current workplace legislation has so far worked, but there needed to be standards to make sure AI was secure.

“There haven’t been many instances where a different set of regulation would have been helpful … so it’s currently all managed through conventional legislation,” he said.

“What we don’t have is a reliable system that ensures the technology itself is safe.”