Portal de Periódicos da CAPES

Google Health and the NHS: overcoming the trust deficit

2019; Elsevier BV; Volume: 1; Issue: 8 Linguagem: Inglês

10.1016/s2589-7500(19)30193-1

2589-7500

Jessica Morley, Mariarosaria Taddeo, Luciano Floridi,

Ethics in Clinical Research

On September 18, 2019, DeepMind's health teams—responsible for artificial intelligence (AI) health research—joined Google Health. DeepMind Health was created with the motivation of seeing the UK National Health Service (NHS) thrive, so a similar ambition can be assumed of the UK branch of Google Health. Certainly, the aims of the NHS long-term plan will be more achievable with Google Health technology. However, unlocking these and other opportunities will not be easy, because of a fundamental problem: a deficit of trust. Public trust in Google Health could have declined following the DeepMind Health deal with the Royal Free NHS Foundation Trust, because the UK Information Commissioner's Office ruled the Trust not to comply with data protection law. Furthermore, AI health solutions have yet to make a significant difference to clinicians on the frontline.1Panch T Mattie H Celi LA The "inconvenient truth" about AI in healthcare.NPJ Digit Med. 2019; 2: 77Crossref PubMed Scopus (128) Google Scholar This ineffectiveness, combined with so-called Dr Google allegedly facilitating a rise in what has been termed cyberchondria,2Aiken M Kirwan G Prognoses for diagnoses: medical search online and "cyberchondria".BMC Proc. 2012; 6: P30Google Scholar indicates that clinicians' trust in AI health solutions could also be low. Fortunately, measures to gain trust need not be complicated.3Taddeo M Modelling trust in artificial agents: a first step toward the analysis of e-trust.Minds Mach (Dordr). 2010; 20: 243-257Google Scholar Google Health and the NHS could and should build on existing frameworks and work together to gain and maintain users' trust. First, both organisations should engage regularly with those who could be affected by the introduction of Google Health technologies, and affected individuals should participate in the shaping of the system.4Durante M The democratic governance of information societies: a critique to the theory of stakeholders.Philos Technol. 2015; 28: 11-32Google Scholar Second, we recommend that Google Health re-establish the independent review board that was in place at DeepMind. The board should operate transparently and be responsible for monitoring, analysing, and addressing the normative and overarching issues that arise at the individual, interpersonal, group, institutional, and societal levels in AI. Third, this board should push Google Health and the NHS to publish the aims and scope of all partnerships, and to publicly commit to open evaluation, feedback, and reproducibility of results. Capitalising on the opportunities of Google Health–NHS partnerships will be challenging, but it is possible. If the two organisations follow these steps while policy makers remain mindful of the risks, ready to redress errors, and wary of vendor lock-in, then hypothetical benefits could become reality. The alternative is to risk losing a significant portion of trust—not only in Google Health, but also in AI health solutions—resulting in a stifling of innovation and considerable opportunity costs. Avoiding this risk must be a key priority. JM is employeed by the UK National Health Service X, which had no input into or influence on the submitted work. All other authors declare no competing interests.