AI firm sacks Aussie staff after $32m government grant

A government-backed Australian health technology company has cut a significant number of local jobs while simultaneously launching a United States affiliate that experts say raises serious conflict-of-interest questions.

Harrison AI, founded in 2018 by two brothers, develops artificial intelligence software designed to assist radiologists in interpreting medical scans. The company claims more than 3,500 practitioners across 1,000 locations worldwide use its platform, and in early 2025 it secured hundreds of millions of dollars in funding — including a $32 million injection from the federal government's National Reconstruction Fund Corporation — pushing its valuation into the hundreds of millions.

Despite that funding windfall, the Sydney-based company moved to restructure its workforce, shedding staff in a move it framed as a pivot toward an AI-first operating model. Multiple former employees posted publicly about their departures, and sources familiar with the company indicate the number of roles cut runs well into the double digits. Separately, some staff reportedly chose to leave voluntarily, disagreeing with the company's new direction — one in which workers would increasingly oversee AI agents rather than perform tasks directly.

The layoffs coincided with the quiet launch of a US-based affiliate, Frontier Radiology, incorporated in Delaware. Unlike Harrison AI — which supplies the tools radiologists use — Frontier Radiology directly employs doctors to perform teleradiology services using those same tools. In other words, Harrison AI has moved from being purely a technology vendor to also being an employer of the medical professionals its technology is designed to assist.

That dual role has attracted scrutiny from health AI researchers. The central concern is how any conflict of interest would be managed when a company both builds the diagnostic tool and employs the clinicians who rely on it. Doctors carry an overriding duty to their patients, which includes the obligation to choose the best available instruments for care — but experts question whether a radiologist on Frontier Radiology's payroll would feel free to challenge, or simply opt not to use, Harrison AI's own software.

Researchers also point to a well-documented phenomenon known as automation bias, where users of AI systems tend to defer to the machine's output even when their own clinical judgment suggests a different conclusion. The concern is that employment ties could compound this bias, making doctors less likely to push back against the tool that their employer created.

Harrison AI's website describes Frontier Radiology as a separate, independent entity that works closely with the parent company. The firm's precise ownership structure is unclear given its Delaware registration. ABC News technology reporter Cam Wilson, who broke the story, said repeated requests for comment went unanswered; the company's co-founder briefly took a call before hanging up.

The company is no stranger to controversy. In 2024, reporting revealed that Harrison AI had built its algorithms using millions of medical scans belonging to Australians — obtained through iMed, the country's largest radiology provider and a major investor in the company — without the knowledge or consent of patients. The Privacy Commissioner investigated both organisations but cleared them on the basis that the images had been de-identified.

Harrison AI maintains that its platform reduces radiologist error rates and boosts efficiency. Proponents argue that a well-run teleradiology service could meaningfully expand access to specialist imaging, particularly in underserved areas. But health experts say those potential benefits need to be weighed against robust governance structures — ones that, as yet, the company has not publicly detailed.