0.4 A Note on Terminology

It can be challenging to find clear terminology in an area of rapid technological change.

‘Algorithmic and data-driven technologies’ will be used to cover diverse technologies that use contemporary computer processing to analyse large amounts of data algorithmically. This includes technology variously described as ‘artificial intelligence’ (AI), ‘machine learning’, ‘neural networks’, predictive analytics, ‘deep learning’, natural language processing, robotics, speech processing and other forms of automation, that are used for the purposes of making decisions.¹⁶

Other data-driven technologies that don’t explicitly use contemporary algorithmic technology – such as electronic records management software, online-counselling platforms, and even some forms of machine learning – remain relevant, as they form part of a broader communication ecosystem that can generate and transmit data concerning mental health.

We use the term ‘communication ecosystem’ to refer to the complex, global networks of information and communication technology. This contemporary communications environment encompasses disparate systems – such as the web, the Internet, and various public and private intranets – which are increasingly converging to create massive, complex, and interconnected flows of data. Other technical terms related to specific technologies will be defined as they arise throughout the report.

The aim of the report is to contribute to public governance. ‘Public governance’ includes law and policy but also extends to professional and ethical standards and guidelines, industry norms, civil society advocacy, and cultural expectations around what members of the public consider socially acceptable. Attention to the diverse relationships of power between these mechanisms can help identify the obligations of those employing data-driven technologies, and the rights of those who use and/or are subject to them.

0.5 Minding Language about Mental Health and Technology

Finding appropriate terminology in global discussions about mental health is also challenging. There is no single set of definitions to describe people’s experience of mental health. Indeed notions of ‘mental health’ are contested. Different terms may be preferred according to national and cultural norms, professional conventions, and so on.

‘people with lived experience and psychosocial disability’ will be used in this report to describe people with firsthand experience of mental health services, mental health crises, extreme distress, psychosis, and so on. We have also sought terminology that conveys our intent to a diverse audience, though we acknowledge that language around mental health is often contested.¹⁷

People with lived experience can have varying ways of understanding experiences that are often called ‘mental health conditions’, ‘mental health challenges’ or ‘mental illness’. We acknowledge that mental health can be described using terms such as emotional distress, trauma, mental health crisis and neurodiversity; and that people may describe themselves as ‘service users’, ‘consumers’, ‘psychiatric survivors’, ‘disabled’, ‘ex-patients’ and so on, and others may reject designations of their experiences in the terms of psychiatry and psychology.

The term ‘people with psychosocial disability’ may not be familiar to readers. It is simply used to refer to people with disability related to mental health. The term has become prominent since the UN Convention on the Rights of Persons with Disabilities (‘CRPD’) came into force in 2008, and its definition is crucial for this report. The CRPD establishes that ‘disability’ includes ‘mental impairment’ (Article 1). Importantly, the CRPD also covers people who experience harms due to imputed impairment or disability—that is, where a person is perceived by others to be impaired or disabled. This is important here because some algorithmic technologies purportedly function to deduce the inner states of human beings, including inferring mental health conditions and cognitive impairments. For example, at the time of writing, Apple is working with multinational biotechnology company Biogen and UCLA to explore using sensor data (such as mobility, sleep patterns, swiping patterns and more) to infer mental health and cognitive decline.¹⁸ Regardless of the accuracy of such predictions there remains the very real possibility of harms against people based on such data, again, even if those data are false, misleading or inaccurate.

The wide definition of psychosocial disability also highlights that all persons may interact with systems that generate data concerning their ‘mental health’. If you carry a smartphone into a counselling service, visit a depression information website, write about distress on a social media platform, or even simply type and scroll on a mobile device, then various ‘digital trails’ will be generated that could be used to infer particular mental states—including mental health conditions, distress and cognitive impairment.¹⁹ This is not to endorse or accept such claims but to reiterate that the rise of Big Data and AI has increased the likelihood of inferences and predictions being drawn from the behaviours, preferences, and private lives of individuals. These data have been described variously as ‘emergent health data’,²⁰ or ‘indirect, inferred, and invisible health data’.²¹ These forms of sensitive personal data create new opportunities for discriminatory, biased, and invasive decision-making about individuals and populations.²² High risk applications of technologies are likely to have the greatest impact on people who are traditionally marginalised, such as those who are using mental health services or those who are subject to involuntary mental health interventions, and those with intersecting forms of marginalisation, but the range of technologies discussed in this report raise issues that affect everyone.²³

The term ‘data concerning mental health’ is therefore used throughout this report to mean personal data related to the mental health of a person, including the provision of health care services, which reveal information about his, her, their mental health status.²⁴ Questions remain about what exactly constitutes ‘data concerning mental health’. The rising use of indirect, inferred, and ‘invisible’ health data is blurring the boundary between ‘patient’, ‘service user’, ‘consumer’ and ‘citizen’, creating multiple issues around the commodification and commercialisation of health, the rise of ‘bio-surveillance’, and other issues which have profound implications in the mental health context and beyond.

This report occasionally departs from these key terms when referring to specific data sources, describing research, or quoting an individual or organization, to accurately reflect the views presented in these materials

Stop the Algorithm by Stephanie Kneissl and Max Lackner in Science Gallery Melbourne’s MENTAL. Photo by Alan Weedon.