As artificial intelligence matures as a technology, policymakers are thinking more about how it can be used in the public interest instead of against it.
Leading the charge in Australia is the Commonwealth Scientific and Industrial Research Organisation (CSIRO), and they think they’ve hit a winner in healthcare.
The latest version of the national science agency’s AI Trends in Healthcare report highlights the rapid growth of AI as a tool for diagnosis and treatment.
The report also highlights significant ethical and regulatory challenges ahead if these technologies are going to be harnessed effectively.
CEO and research director of CSIRO’s Australian e-health research centre David Hansen said policymakers need to try to get in front of the commercialisation of technologies if AI is to be harnessed effectively for healthcare.
“A key difference between the use of AI in healthcare and the use of AI in other industries is where AI provides decision making for diagnosis, prevention, prediction, prognosis, monitoring or treatment,” he said.
“In these cases, the AI is considered a medical device and is currently regulated as such — ‘software as a medical device (SaMD)’. This feeds into a bigger discussion of the use of AI across healthcare, for clinical and non-clinical purposes, and ensuring Australia is ready for its use.
“The healthcare consequences of the rise of generative models are rapidly unfolding and the national discussion about how to regulate AI is gaining pace.”
The report examines four key areas of development for AI in healthcare, and the implications for both healthcare consumers and regulators. These tie into areas where AI already has or soon will have the capacity to improve people’s interactions with the health system at minimal risk.
Interoperability is an area people have been looking for healthcare efficiencies for close to a decade.
The ability to create systems that share health data between different practitioners who need access has been hampered by concerns about the government’s cybersecurity capabilities and the manner in which a previous attempt to create a universal government health record was foisted on people with little consultation on the social consequences of such an action.
Ensuring citizens have trust that their health data is being stored and transferred safely and won’t be accessed by people who may it against them (like prospective employers) is fundamental to the introduction of AI in healthcare.
Without personalised data accessible through all health systems, it is unlikely to result in practical improvements in people’s health outcomes.
This would demand the creation of cloud infrastructure to support secure data storage, transfer, and processing by the health sector. Some economists say the only practical way for this to be done is for government to treat essential digital infrastructure as a publicly owned utility.
As we enter the age of personalised healthcare and consumer products, apps and customisation of the presentation and processing of data would need to be available for practitioners and patients.
Finally, data and analytics would need to support all of these systems to provide feedback on what is working, and what isn’t,”
“To fully harness AI and machine learning … we need not to just let it happen, but rather plan for its introduction into healthcare.
“This means we will be able to benefit properly from AI by ensuring the frameworks are firmly in place for ethical implementation and that the safety, quality and monitoring guidelines are established as we strive to create newer and better AI-based digital tools,” said Hansen.
READ MORE:
Medicare report recommends breaking down healthcare data silos
