Google DeepMind and healthcare in an age of algorithms

From Digital Culture & Society

Hodson, Hal., Powles, Julia. (2017). Google DeepMind and healthcare in an age of algorithms. Health and Technology, 7(4), 351-367.

Find article online:

DOI: 10.1007/s12553-017-0179-1

Contents 1 Context 2 Overview 3 Strengths and Weaknesses 4 Assessment

[edit] Context

This article discusses the relationship of DeepMind Technologies Limited and Royal Free London NHS Foundation Trust with their collaboration in the management of acute kidney injury. DeepMind which is apart of the Google conglomerate Alphabet Inc. is a company that specializes in artificial intelligence (AI). The purpose of this article was to analyze the case study of DeepMind-Royal Free by evaluating datasets on large private prospectors and identifying critical questions within healthcare. This case study was performed when Royal Free transferred the records of patients without consent in order for the development of a clinical alert app for kidney injury.

[edit] Overview

This article is split between three sections. The first two sections discuss how DeepMind was able to gather data for Royal Free by giving a chronological timeline of the case study. The first section discusses the contractual work performed between DeepMind and Royal Free and also explains the algorithm used to detect acute kidney injury (Hodson p. 353). The authors point out how the algorithm has the probability of failure when attempting to detect acute kidney injury. The algorithm of detecting acute kidney injury provides a way to detect any potential harm in patients through the history of their medical data. DeepMind had developed a medical app as the function in detecting acute kidney injury. In the second section, the process of the patient data being given without consent is highlighted where the legality of the issue is discussed (Hodson p. 354). In this section the authors highlight how data-driven tools should be grappled within the domain of public healthcare. The third section explores DeepMind’s ambitions in their reasoning of their relationship with Royal Free. The authors elaborate of how the transfer of data is a problematic issue that may be abused in the future (Hodson p. 358). In this section there is elaborate discussion about the protection of data and how there is a responsibility for companies to keep privatized data in a secure fashion.

[edit] Strengths and Weaknesses

The main strength of this article is how it expresses how algorithmical searches of databases can be a concern, especially due to how automation and machine learning become more prevalent technologies that are utilized. The article does a brief but concise explanation as to how algorithmic data can be utilized in order to alert medical practitioners of acute kidney injury in their patients, expressing how machine learning can play a pivotal role in the realm of healthcare. The second strength is how the authors express their concern of subsidized data is an issue, although this is done extensively within the article. It is important to understand how there are limitations and challenges when large datasets are being utilized as there is the possibility of data that is not related towards the purpose of the data being breached.

The main weakness of this article is how it is focused upon the privacy concerns of the relationship between DeepMind and Royal Free while not going into great detail about how the algorithms pertaining to detecting acute kidney injury are utilized. This lack of explanation comes across as a bias of the authors to try and explain why algorithms and the transfer of data between governments is an issue. Outside of acknowledging the relationship the case study between DeepMind and Royal Free the authors do not reference other practices between companies from a statistical sense. There is a narrow-minded focus on explaining why the transfer of data is an issue of concern.

[edit] Assessment

In conclusion, this article provides a light to the concerns of practices that companies such as DeepMind perform when subsidizing data from other companies that have rich databases such as Royal Free. Although the algorithm that was developed can help detect acute kidney injuries in their patients, there are many privacy concerns created when data is being traded between companies. Machine learning is a tool that can help physicians in their prognosis of diseases such as acute kidney injury. While the authors focus on the morality of data privatization, health datasets can be utilized in order to detect problems that can be present in potential patients. Nonetheless, this article addresses legitimate concern to how subsets of data can be abused and how the transfer of data between companies for the usage of algorithmic machine learning can be a hindrance to the privacy of patients.

---

--Bl14hl 22:11, 19 March 2019 (EDT)