The human advantage in an age of human-like AI

Are humanoids workers or tools?

China is a global leader in humanoid robotics, but we’re already seeing this tech in Australian settings. One of its most significant applications isn’t as a runner, but through an emerging technology known as empathetic humanoid companionship.

During COVID-19 lockdowns, then 21-year-old Grace Brown built her first aged care robot in her Melbourne bedroom. Named “Abi”, this machine was the first of 22 robots now deployed to NFP provider Mecwacare. The Abis are 1.2m tall and speak up to 90 languages. According to Brown in a Sydney Morning Herald interview, they occasionally blow bubbles and remember “details such as the songs residents grew up with”.

As friends of the elderly, the robots make a lot of sense. They can provide personalised conversations based on past interactions, facial recognition and AI learning models. They can sing, dance and share fond memories. The Abis are true companions, not substitutes for aged care workers or medical staff, and are brightly coloured with cartoon eyes and geometric limbs.

However, in South East Asia and the US, some of these devices – and they are devices, Moretti encourages us to remember – speak in comforting tones from human faces.

In this setting, the humanness of the robot is the point, designed to mimic support services and offer familiarity and comfort. Even the language becomes tricky. Are they “working in” aged care facilities? Perhaps “operating at”? Surely not “living in”?

The rise in humanoid companions coincides with skyrocketing cases of AI love. In the US, one in five high-schoolers has had a romantic relationship with AI or knows someone who has (Center for Democracy and Technology).

“People are easily slipping into a kind of psychosis,” says Moretti. “This is addictive technology. How do we ensure we’re thinking about the impacts it has?”

So convincing are these bots and apps that China has now established the world’s first national standard to clearly separate man and machine. Under these regulations, AI must – among other requirements – clearly and repeatedly disclose it is not human. It must detect and intervene in emotional human distress. It cannot simulate relatives or specific family members in order to interact with the elderly. And it must have ethics review mechanisms built in.

Are our regulations ready?

By contrast, despite strong public support for AI regulation, Australia has no such laws.

“It’s a very fragmented space for an organisation to be able to say, there’s a uniform standard on what this looks like,” says Moretti. “It’s difficult to apply technical and regulatory standards. But AI hasn’t emerged into a regulatory vacuum. Consumer product law, competition law and privacy laws all still apply. Just because AI is a new technology, organisations sometimes feel these are no longer a part of the question.”

But the lack of specific AI regulation, for now, is no excuse not to be thinking about it. “Organisations need to be mindful about how they move forward on what will become a compliance issue, even though it’s purely an ethical one now,” says Moretti.

Organisations can be adapting processes and frameworks they already have in place, she says, and prioritise the human experience as much as possible.

“When we think about technology and risk, we think about vulnerabilities in the technical system. I’d strongly encourage organisations to look outward to people. What is the impact this might have on a person, on a community of people, on the environment? Start thinking about building out those risk registers.”

According to Deloitte research, 70 per cent of healthcare professionals believe AI will “significantly transform the industry within the next five years”. It’s already used for everything from performing surgery to dispensing medicine (and blowing bubbles). But we turn to humans to deliver difficult news or notice a change in a patient’s demeanour, and only four per cent of Australian employees think AI can interpret health data without human intervention.

The healthcare industry represents one of Australia’s greatest opportunities for AI adoption. But tech and humans must work harmoniously to be successful. We must consider the ethical implications and plan to operate within a regulated space, ready to innovate, but not devastate. And, says Moretti, we must slow down.

“You can be fast or you can be something else. Do you want to be fast, or do you want to be good? Do you want to be fast or fair? If efficiency is front and centre of your AI strategy, you’ll trip over a lot of other human values along the way.”

This article first appeared as 'The heart of the matter' in the June/July 2026 Issue of Company Director Magazine.