Much has been written on the introduction of Artificial Intelligence (AI) into many aspects of our lives, our work and our leisure.
There has been an increasing number of articles and reviews of the use, or potential use, of AI in health.
Many advocate for the use of AI for the obvious advantages and improvements that can occur in treatment, diagnosis and care.
Some articles warn of the dangers of an untried and untested technology, and which may be fraught with errors and miscalculations, given its early stage of development. There is also a risk of overstating the benefits of AI at this early stage.
Given that so much in medicine is based on the “precautionary principle”, we should look forward to exploring the great advances that AI may present in the health sector, but also be alert and wary, in moving too fast before appropriate testing and confirmation has occurred. As part of recognising that risk, there is an inherent legal risk in a swift adoption of AI without appropriate testing, confirmation and recognition of clear benefits.
These legal risks include:
Negligence
Untested and unconfirmed AI applications could lead to results which harm patients, and expose users to claims of negligence. Raw acceptance of data generated by AI or solutions created by AI, without adequate or independent confirmation by the user, could clearly frame a claim in negligence at law.
Privacy
AI depends on significant amounts of data by which it formulates outcomes, and in health, much of the data relates to patient information, where patient consent is required. Institutions will need to ensure that their existing privacy statements permit, or new privacy statements are created which permit, the extraction of data for inclusion in AI applications. Such data should, to the greatest extent possible, be de-identified, so as to preserve the privacy and confidentiality of patients.
Notice and Consent
Particularly in the early stages of development of AI applications in health, patients might need to be put on notice, be aware and consent to the use of treatments, prescriptions or applications which are based on AI predictive modelling or untried and untested developments. Patients can accept known risks, which are adequately explained and where the potential harm arising is explicitly consented to and accepted by them. A failure to alert patients to treatments based on AI may also form the basis of legal claims.
Insurance
The ultimate fall back for professionals and hospitals in the event of claims, is insurance cover – both to protect the healthcare provider and compensate an injured patient. Not all insurance policies will cover the novel use of AI developed care and treatments. A review of insurances should be undertake to identify the extent (or lack) of cover.
In one sense, these challenges are not new. Medicine has developed, sometimes quickly, with the advance of knowledge, technology, information and research. The introduction of robots in surgery was initially met with scepticism, but, after appropriate testing and checking, robots are now a regular feature of some elements of surgery. Just as robots are not the best solutions for all types of surgery, it will be expected that the application of AI will not necessarily enhance all aspects of health and patient treatment. No doubt there will be testing times through the initial development and enhancement of AI, before these matters become clearer.
Just as with any new drug or treatment, the medical profession would undertake appropriate trials to better explore and determine the efficacy of them. The same approach should be considered in relation to advances in the use of AI applications, particularly where impact on patients is involved.
Like with the introduction of other novel treatments and applications in the past, caution, awareness and confirmation of outcomes are the “friend” of health providers.
How we can help
If you have questions or require further assistance in relation to AI and health law, please contact Michael Gorton AM, Jonathan Teh, Gina Tresidder or a member of our Health team.
If you would like to stay up-to-date with Russell Kennedy’s Alerts and Insights from our Health team, you can subscribe to our mailing list here.