A leaked e mail from tech firm harrison.ai to buyers reveals an govt blaming the radiology chain I-MED for considerations raised about utilizing non-public medical knowledge to coach AI with out the sufferers’ information, because the start-up provides its first public defence of its practices.
A Crikey investigation revealed the buzzy healthcare firm had skilled its flagship AI product on doubtlessly a whole lot of hundreds of Australians’ scans obtained from the nation’s largest medical imaging supplier seemingly with out categorical consent from its sufferers.
Now Crikey has obtained an e mail despatched by harrison.ai chief working officer Peter Huynh looking for to settle buyers spooked within the fallout of the investigation.
Despatched at 8.39 pm on Thursday night time, Huynh defended his firm, shifted duty over affected person consent to I-MED and made new claims about its knowledge practices.
“For readability, the authorized foundation for utilization of I-MED knowledge by Harrison is by way of the info licence settlement that we’ve in place, ruled by related legal guidelines,” he wrote.
Huynh rejected any position within the knowledge’s provenance, shifting duty wholly to the harrison.ai accomplice and shareholder: “Questions posed by the article in relation to I-MED knowledge assortment and consent will not be issues for Harrison to answer.”
He added that the corporate had obtained ethics approval for its medical research and complied with the legislation.
Earlier than signing off, he requested in underlined textual content for buyers to “chorus from remark” and ahead any media request to harrison.ai. Huynh didn’t reply to Crikey’s calls or messages.
On Friday afternoon, harrison.ai gave its first public acknowledgement of the controversy in a press release, supplied to Crikey, ignoring particular questions on its conduct.
“As a clinician-led firm, harrison.ai takes affected person security and knowledge privateness very severely. Our merchandise are enhancing high quality of take care of sufferers world wide by guaranteeing earlier and extra correct detection of ailments,” it stated.
“The information we obtain for analysis and growth is de-identified, can’t be re-identified, and is encrypted. Such medical analysis and growth is finished in compliance with all related regulation, together with privateness legal guidelines.”
The corporate didn’t present a replica of its ethics approval functions or approvals, which lists how consent was sought or waived for its experiment.
I-MED stays a big shareholder of harrison.ai, and its CEO Clare Battellino is likely one of the tech firm’s seven board members. I-MED’s nationwide communication supervisor didn’t reply to a request for remark.
On Friday, authorities and Greens politicians voiced their concern over privateness points about AI and using healthcare knowledge in response to Crikey’s revelations.
