On Thursday 6th September, LML co-hosted a workshop titled 'Regulating Algorithms in Healthcare: Intellectual Property and Liability' with the PHG Foundation and the Center for Advanced Studies in Biomedical Innovation Law. The workshop brought together a multidisciplinary group of academics, researchers, policy makers and regulators, to evaluate intellectual property and liability issues arising from use of algorithms and software in healthcare. This is an area of significant interest to LML, as Dr Jeff Skopek (LML Deputy Director) was recently awarded a grant to study liability for uses for AI in healthcare by the Wellcome Trust. LML was represented by Kathy Liddell, Jeff Skopek, John Liddicoat, Mateo Aboy, Cristina Crespo and Matt Jordan.
The first half of the workshop was dedicated to discussion of intellectual property protection for algorithms in healthcare. Prof. Mateo Aboy (Senior Research Scholar with LML) gave a presentation on the patentability of algorithms in the healthcare space and new empirical evidence about the impact of US cases such as Myriad, Mayo and Alice. Dr Enrico Bonadio discussed the relationship between free and open source software (FOSS) and IP.
Dr Kathy Liddell (Director of LML) facilitated the IP panel, framing discussion in terms of identifying and refining what might be considered to be IP policy problems, assessing the size of any problems, and considering what the responses might be. Prof. Aboy and Dr Bonadio were joined by Prof. Timo Minssen (CeBIL), Iain Mitchell QC, and Andrew Katz (Moorcrofts). Several cross-cutting questions and themes emerged during the morning's session. For example, to what extent are IP policies in this landscape pitted against each other and to what extent are they complementary and capable of being integrated? Are the patent law issues primarily legal uncertainty, or are there problems with legal principles as well? Supposing it was possible, would it actually be desirable to increase the use of FOSS in this space as an alternative to IP protection? What are some of the complexities of 'free' and 'open' innovation systems?
The second half of the workshop concerned possible models of liability for algorithms where harm occurs as a result of their outputs. Dr Jeff Skopek introduced Professor Glenn Cohen (Harvard Law School) and Professor Nicholson Price (Michigan Law School). Prof. Cohen gave an account of the increasing role of AI and predictive analytics in healthcare, before focusing on tort law as lens through which one might debate the optimal scheme of liability, and how liability might be apportioned to the relevant actors. Prof. Price presented on medical malpractice and black box medicine, posing the question of how schemes of negligence might set the incentives for bringing algorithms into practice.
Joining the liability session speakers for a panel discussion (facilitated by Alison Hall, Head of Humanities at the PHG Foundation) were Dr Alberto Gutierrez (former Director of the FDA's Office of In Vitro Diagnostic Device Evaluation and Safety), Dr Danielle Belgrave (Microsoft) and John Buyers (Osborne Clarke). They addressed questions of liability from the viewpoint of regulators, researchers, and lawyers respectively.
This event was the second in a series of two workshops held by the PHG Foundation on the topic of regulating algorithms in healthcare. The first focussed on the GDPR and IVDR in practice. The PHG Foundation plans to publish reports on both events.