ARIA Research’s augmented reality eyewear promises to give vision-impaired people more freedom than ever. Will the public concern over AI-enhanced and camera-equipped glasses hold them back?
The more the world sees of smart glasses, the less it likes. The AI-enhanced, camera-equipped spectacles have been used to harass retail workers, intrude upon women in public spaces and capture footage of people going to the bathroom or changing clothes.
It’s an uphill public relations battle that Australian smart glasses pioneer ARIA Research has been ready for. In fact, the company’s CEO and co-founder Robert Yearsley told create, these problems are something his team has been thinking about for half a decade.
“I’m a privacy advocate,” Yearsley said. “That’s probably interesting to point out.”
READ: Meet the company helping the visually impaired see through sound
Developing a disability aid
ARIA’s glasses are specifically designed as a disability aid: using AI, machine learning and audio feedback, the glasses are designed to help blind and vision-impaired people navigate the world around them.
“The idea behind ARIA was to map visual spatial perception into a format that someone without vision could understand through sound,” create wrote when the company first appeared in this publication two years ago.
“ARIA’s system uses machine vision to map the environment, build a three-dimensional virtual model, immerse the user in that computer-generated space, and render the sound of objects in their actual locations, in real time.”
Since then, Yearsley has trialled the glasses with 12 vision-impaired users to validate the safety and usability of the technology. Currently he’s pushing towards an actual product launch. But the privacy questions have remained on his mind.
“The prediction back then was it’s likely that Meta is going to, for better or worse, drive this conversation, and it certainly turned out that way,” he said.
“I’m in an interesting position where I’m talking about cameras on someone’s face that are videoing all day, but we’re not recording the video to pass on – to observe.
“We’re trying to turn it into metadata to provide empirical real time benefit. It’s a subtle distinction but a really important one.”
READ: 3 innovations transforming biomedical engineering
User first
And that benefit could be a remarkable one. According to Yearsley, 338 million people around the world have a vision disability, and the best solutions available at the moment are “canes, guide dogs, and human assistance”.
“The degree of agency and independence afforded by internet-connected, AI, machine-vision systems are capable of providing is just extraordinary,” he said.
“It would be a damn shame to have that swept up and lost as a result of [policymakers] not really thinking about the motivations and use-cases of these technologies.”
Policy tends to lag technological innovation, Yearsley said, meaning it’s often a company’s business model that shapes its approach to privacy. And there are opportunities for companies to do the right thing on privacy not because policy has been put in place, but because it offers an intrinsic advantage for the product and its users.
“Who’s our customer?” he asked. “It’s the actual end user – it’s the blind user. What happens with data? We should respect it – it’s their data. What happens if we want to access it to improve our product? Shouldn’t we have a separate arrangement where we pay them for that data?
“We’re in a position to look at ideas like this, and that is largely driven by the funding and benefit and business model.”
So when it comes to smart glasses technology, Yearsley urges policymakers to “always consider the source” and think about unintended consequences.
“If you imagine the development of smart glasses, then AR systems, and the turbocharging of generative AI and LLMs and VLM models, think about the businesses and where that’s coming from,” he said.
“Meta obviously has a very specific commercial interest, which is that the user is the product. And that definitely influences design decisions thereon – the kind of businesses, consumer offerings and decisions that are made.
“Likewise, for us, our business model and our benefit have a completely different set of motivations, which is ‘How do we help a blind person have spatial perception?’”
READ: These Australian companies are building the country’s biomedical capacity
A different perspective
ARIA is developing a medical device rather than a consumer good, Yearsley said, which affects its perspective – as well as its constraints – about safety, privacy and how it treats the information it gathers.
“We don’t actually need to see all this information,” he said. “We saw some opportunities, from an architectural engineering perspective, to embody some of these ideas [by] running machine vision models and data analysis on-device, for instance.”
ARIA’s use case makes a big difference too, he said. The smart glasses his company produces gather visual information and metadata to provide the user with audio cues so they can provide a spatial interpretation of the environment.
“But it’s the same underlying technologies, so our worry and concern is that policy basically sweeps everything up at the same time without thinking about some of these not just edge cases,” he said.
“I believe we’re working on one of the most beneficial use cases there is for this kind of technology at the moment. That’ll change somewhat as it matures into consumer markets but I’m really hard pressed to think of a higher-value human benefit for a combination of smart glasses cameras and AI.”
