A leaked email from tech company harrison.ai to investors shows an executive blaming the radiology chain I-MED for concerns raised about using private medical data to train AI without the patients’ knowledge, as the start-up gives its first public defence of its practices.
A Crikey investigation revealed the buzzy healthcare company had trained its flagship AI product on potentially hundreds of thousands of Australians’ scans obtained from the country’s biggest medical imaging provider seemingly without express consent from its patients.
Now Crikey has obtained an email sent by harrison.ai chief operating officer Peter Huynh seeking to settle investors spooked in the fallout of the investigation.
Sent at 8.39 pm on Thursday night, Huynh defended his company, shifted responsibility over patient consent to I-MED and made new claims about its data practices.
“For clarity, the legal basis for usage of I-MED data by Harrison is via the data licence agreement that we have in place, governed by relevant laws,” he wrote.
Huynh rejected any role in the data’s provenance, shifting responsibility wholly to the harrison.ai partner and shareholder: “Questions posed by the article in relation to I-MED data collection and consent are not matters for Harrison to respond to.”
He added that the company had obtained ethics approval for its clinical studies and complied with the law.
Before signing off, he asked in underlined text for investors to “refrain from comment” and forward any media request to harrison.ai. Huynh did not respond to Crikey’s calls or messages.
On Friday afternoon, harrison.ai gave its first public acknowledgement of the controversy in a statement, provided to Crikey, ignoring specific questions about its conduct.
“As a clinician-led company, harrison.ai takes patient safety and data privacy very seriously. Our products are improving quality of care for patients around the world by ensuring earlier and more accurate detection of diseases,” it said.
“The data we receive for research and development is de-identified, cannot be re-identified, and is encrypted. Such medical research and development is done in compliance with all relevant regulation, including privacy laws.”
The company did not provide a copy of its ethics approval applications or approvals, which lists how consent was sought or waived for its experiment.
I-MED remains a significant shareholder of harrison.ai, and its CEO Clare Battellino is one of the tech company’s seven board members. I-MED’s national communication manager did not respond to a request for comment.
On Friday, government and Greens politicians voiced their concern over privacy issues about AI and the use of healthcare data in response to Crikey’s revelations.
