Automation Anxiety

From self-driving cars, through high-frequency trading to military drones and organised swarms of shelf-stacking robots, our era is marked by rising automation and a new concern with the likely social, cultural, and economic implications of this computationally driven transformation. This research network will explore innovative methods by which the humanities might address contemporary cultural anxiety about new forms of automation, which we are calling automation anxiety. The network will organise contemporary automation anxiety into three themes which form the basis of each workshop:

1. Human obsolescence: the automation of cognitive labour, and the ends of human expertise;
2. Human (in)security: automated military engagement, law enforcement, and surveillance
3. Human in(attention): automation or delegation to machines which produces anxiety or instability.

Concern with the mechanisation of tasks formerly done by humans extends back to the beginning of the industrial era. The loss of jobs to machines, which John Maynard Keynes called 'technological unemployment', was a recurring economic and social anxiety throughout the twentieth century. There are long-standing economic, historical and sociological debates about the impact of automation. However the current wave of computational technology, using algorithms, machine learning and other techniques, extends automation beyond the factory and into roles requiring complex decision-making, previously an exclusively human realm. While in many cases the computational techniques are not especially novel, the application of these approaches to concrete systems and products has generated a new wave of anxiety about their impact on human culture and society. These concerns manifest themselves in public culture through news stories, editorials, films and airport bestsellers around the notions of, for example, machine learning and artificial intelligence.

The humanities need new tools to analyse contemporary automation anxiety. Much of the debate in public culture expresses itself online and in social media, which requires the exploration of digital methods, such as web scraping, data visualisation, sentiment analysis, and digital ethnography. Equally, an analysis of automation anxiety calls for the humanities to have a better understanding of the computational culture of automation itself. Many concerns arise from the opacity of the technology and the replacement of more-or-less understandable and accountable systems of human decision-making with the 'black boxes' of algorithms or machine learning. For these reasons, research on contemporary automation anxiety calls for a network that is both radically transdisciplinary and focused on research methods.

The proposed network will represent an unparalleled opportunity to bring together a wide range of knowledge and expertise from diverse disciplines and backgrounds in order to generate new collaborative research projects in this area. The workshops will convene scholars in philosophy, language and literature, history, media and communication, cultural geography, journalism, cultural studies, law, history, computer science, economics and sociology together with non-academic users in the fields of artificial intelligence, policy consultancy, innovation advocacy and social media analysis. They will explore dimensions of contemporary automation anxiety including the philosophy of autonomous killing machines, the image of the drone in literature and popular culture, algorithmic law enforcement through computerised 'predictive policing', feminism and the automation of labour, algorithmic governmentality and automation as disruption of the state. The development of UK research in this area will be enhanced and shaped by the inclusion within the workshops of recognised, world-leading experts in the cultural and social impact of automation.

This network will examine the growing human implications of computationally-driven automation. It will produce new thinking and engage with a range of constituencies both affected by and leading policy and civil discussion around automation. The workshops bring academics in the disciplines of philosophy, language and literature, history, media and communication, cultural geography, journalism, cultural studies, law, history, computer science, economics and sociology together with research users in artificial intelligence, innovation advocacy, policy consultancy and social media analysis to discuss cultural anxiety associated with automation. The network will examine methods and digital tools that could be used to analyse that anxiety for purposes that include producing tangible outputs in the areas of social policy and design of services and products. The network will affect the following impact areas:

1. Public understanding of concerns around automation

There is an urgent need to enhance the public understanding of ubiquitous computation of which automation anxiety is a symptom. These include concerns about the implication of automation for the world of work, automated technologies for defence and law enforcement such as drones or predictive policing and human loss of skill and interest due to the delegation of tasks to computational automation. Channel 4 Economics Editor Paul Mason, a confirmed network participant, has demonstrated the importance of academic research around automation to public thinking in his recent "Postcapitalism a guide to our future".

2. Digital tools for analysing public culture

This network's focus on innovation in research methods will benefit those who use digital tools to analyse public culture, such as think tanks, pressure groups and social media analysis companies. Network participants will explore new techniques and methodologies for analysing cultural anxiety; once defined and presented these will be available for academic and non-academic use by contributors to the network such as Brandwatch who specialise in social media analysis. Equally the Sussex Text Analytics Group, working with Demos and Ipsos MORI, have extensive experience of bringing the benefits of text analysis to wider debates in public culture.

3. Management and analysis of public data

The use of algorithms and machine learning to analyse patterns in large datasets has far reaching implications for the use of public data. The gathering, management and analysis of this data will benefit from the network's analysis of cultural anxiety about, for example, forms of law enforcement that target policing based on computational analysis of crime, housing and census data. It will inform policy and decision making about how and why data is gathered and increase understanding about why these anxieties exist.

4. Understanding and developing technological innovation

Public concern about automation has extensive significance for organisations and companies engaged in promoting innovation in the UK. Those involved with developing new forms of automation and managing and promoting innovation in this area, such as Nesta, will benefit from a clearer analysis of public concerns and anxieties about new computerised forms of automation. Groups which engage with technological innovation need to understand better how automation anxieties are symptomatic of wider cultural concerns.

5. Development and design of automated systems

Understanding the human experience of automation from a cultural perspective is of crucial relevance to the successful adoption of design and development of future automated systems.