ETHOS Issue 18 Jan 2018
According to Klaus Schwab, Founder of the World Economic Forum, we “stand on the brink of a technological revolution that will fundamentally alter the way we live, work and relate to one another.”1 The Fourth Industrial Revolution will see the fusion of technologies in which the lines between the physical, digital and biological spheres are blurred. Industry 4.0 will be characterised by billions of people staying connected remotely through pervasive mobile devices, able to access vast amounts of newly created data, while assisted by capable machines.
The adoption of artificial intelligence (AI) is at an inflection point. The internet of things, fuelled by neural-based machine learning, will move from simple digitisation to a convergence of technologies.
The major challenge is not technical but how to integrate robotic automation into workplace processes, decision-making and culture.
Whether this transformation will be good or bad for citizens is hotly debated. Platforms “will increasingly enable citizens to engage with governments, voice their opinions, coordinate their efforts, and even circumvent the supervision of public authorities”.2 Governments will face increasing pressure to engage with citizens whose access to information will be more immediate than ever before. Digital opportunities for democratic engagement will increase, while the organisation of protests or online political advocacy will become easier. Government open-sourcing has the potential to underpin participatory government in a data-driven society, by allowing the public real opportunity to contribute to the development of public policy.3 However, as jobs disappear, the governance of citizens may also become more difficult. Societies may fragment as inequality increases.
New Technologies Will Reshape Public Sector Work
Unquestionably, a new wave of disruptive automation is already transforming work. In the last century, automation saw repetitive lower-skilled factory labour progressively taken over by mechanical robots. Now it is administrative skills that are most under challenge. Much of the expertise of professionals will be demystified and their role as gatekeepers to knowledge undermined.4 Many occupations which previously required extensive training and the exercise of judgement can already be undertaken faster, better and at lower cost by increasingly capable machine-based systems. A high proportion of these jobs are located in public services.
As the technologies of visual recognition and speech translation combine with exponential growth in data processing capability, officialdom is likely to be changed in fundamental ways. It will not just be the relatively routine but often complex tasks of data collection, clerical processing and compliance checking that will become automated. Professional skills, including risk assessment and project management, will increasingly be undertaken with greater reliability by robo-advisors.
The application of robotic process automation (RPA) to administrative jobs will enable significant reductions in the size of public services. Estimating that the US Federal Government could potentially save billions of dollars through automation, Deloitte concludes that “cognitive technologies will eventually fundamentally change how government works, and the changes will come much sooner than many think.”5 Similarly, the UK think tank, Reform, suggests that the UK public sector can be slashed through automation, and not just clerical back-office jobs. Sophisticated “chat bots” that conduct human-like conversations by text or voice will become the new frontline. “Public services” they argue, “can become the next Uber.”6
Much of the expertise of professionals will be demystified and their role as gatekeepers to knowledge undermined. Many occupations which previously required extensive training and the exercise of judgement can already be undertaken faster, better and at lower cost by increasingly capable machine-based systems. A high proportion of these jobs are located in public services.
At the basic level of the public sector workforce, change is likely to be intense. Driven by value-for-money considerations, a desire to lift productivity and an ambition to improve customer service standards, RPA can rapidly enhance the performance of public administration. AI-augmented government will be able to deploy computers to perform tasks and make decisions previously thought to require intellectual effort. They will not necessarily replicate the thinking process of human specialists.
"I want people to use digital public services as readily and as constantly as they do when shopping, socialising or checking bus times. By doing so, we can actually change the way citizens interact with us."7
—John Manzoni, Chief Executive, UK Civil Service
There are dissenting voices. Some remember the false dawn of workplace automation in the early 1980s, when many observers anticipated a workless society that never came to be. Not a few are jaundiced by the disappointment of too many earlier ‘e-government’ initiatives and remain sceptical of rhetoric. The former Chief Executive of Australia’s Digital Transformation Agency, Paul Shetler, has criticised how public service leaders tend to talk up a bold future of “big data, artificial intelligence, dancing holograms, and all that kind of stuff” yet persistently fail to deliver on the basics of customer service.8
Too often, government agencies imagine the future without effectively delivering the present. Too frequently their application of technology simply reproduces the traditional hierarchical structure and functional demarcations of the public service.
Bureaucratic Processes Lend Themselves to Automation
Much public administration is ideally suited to robotic application. Bureaucracy is a rules-based system, governed by legislative and administrative guidelines, comprising many routine procedures for decision-making. Today, these repetitive processes require public administrators to exhibit expert knowledge of complex systems and arcane criteria in order to make decisions. Tomorrow, most of these tasks will be able to be undertaken by machines. Providing information, making payments, collecting revenues, processing grant applications and enforcing regulations will increasingly be automated. So, too, will government procurement, property maintenance, payroll processes, IT diagnostics and the provision of shared services.
In most instances, robotic administrators will be able to undertake such tasks faster. Working 24 hours a day, every day of the year, backlogs will be reduced. The contingent liabilities associated with paid labour will be lowered—machines do not need recreational or medical leave (unless the IT system goes down), or complain about workplace conditions.
Public sector performance will be lifted. Quality will be improved. The robot that emulates human execution of repetitive tasks will make more accurate decisions. It is less likely to make errors or be swayed to exercise discretionary judgement in uncertain ways. Public accountability will be enhanced. It will be easier to audit every robotic action and decision. Compliance and security requirements will be consistently applied. As machine-based speeds and volumes increase, the scalability of technological innovation will increase vastly. At present it is civil servants who provide the glue that joins together the complex network of systems, processes and rules that underpin effective public administration. In the future, it will be robots.
Computers, without explicit programming, will learn how to learn, mining unstructured information to detect patterns to anticipate events. Robots are already developing better-than-human predictive capabilities, able to identify both patterns and anomalies in huge data sets. They will be able to foresee the spread and direction of health epidemics, analyse patterns of transport use or identify the likelihood of adverse weather events. They will be able to spot irregularities that might indicate welfare fraud, insider trading, identify theft or money laundering.
Robots will have the capacity not just to repeat processes but to discern methods of improving them.
Consumer and citizen online behaviours are DIFFERENT
A citizen in search of public services is not a shopper. Assumptions born of consumer behaviour in a competitive market economy do not readily translate into the relations between an individual and the state. There are important differences.
First, in most instances citizens do not choose to do business with government. It represents an unavoidable impost on their time. Looking online for the best value purchase can be an enjoyable experience. Submitting one’s annual tax return, paying a fine for speeding, applying for a family payment, are irritating chores. Most citizens want to complete their transaction as speedily as possible, at a time and place of their convenience, with easy access to assistance if the matter turns out to be more complex than anticipated (it often does). That sense of government as a necessary burden should drive the initial application of automation and digital communication.
Perhaps with time, perceptions might change. Citizens might be persuaded to use automated information technology to increase the range of things they actually want to do with government, such as booking a medical appointment online or reporting a pothole in the road. Making publicly-funded data available to the public under a creative commons license will allow them to use government-collected information in ways they find useful: for example, to check out house prices, peruse land titles or identify which suburbs are safest to live in.
Second, citizens appear to hold a view of privacy that is different from that of consumers. Young people, in particular, seem to be increasingly cavalier about the personal details that they put up for others to read or view online. They are remarkably sanguine about placing information about themselves on social media. In contrast, citizens remain wary of the capacity of governments to hold their confidential information securely. Public trust in governments and politicians tends to be low and declining. Citizens demand public protection of their private lives at a level that they often fail to require of each other in the world of online ‘friends’. They distinguish between the perceived freedom of the market and the executive authority wielded by the state.
These important caveats may constrain the movement of automation from the private to the public realms.
AI-Assistance Will Move from the Back-Office to the Frontline
As cognitive technologies are applied—through computer vision, speech recognition, natural language processing and machine translation—robots will move from the back-office to the frontline. It’s already happening. Apple’s Siri and Amazon’s Alexa now have public service colleagues working as voice-driven virtual androids. The US Immigration Service introduced a virtual assistant, Emma, to help answer 14 million enquiries a year. To assure privacy and confidentiality, Emma promises to delete chat logs at the end of each session.9
Robotic automation widens the opportunity to apply behavioural psychology to “nudge” citizens towards pro-social conduct. Machines can be programmed to engage with members of the public in different ways, learning what approaches have most positive impact. They can test responses. Whether it is persuading recalcitrant offenders to pay their fines or convincing families to live healthier lifestyles, robots can learn how to communicate in the most effective manner. They can teach us what works best with humans.
Much public administration is ideally suited to robotic application.
All of this has the potential to add significant value to the manifold activities of government that will continue to require a human presence. Converging technologies will allow people to work together in quite different ways. Much of the delivery of public value is already undertaken in partnership with the private sector or in collaboration with community-based organisations. Nevertheless, the institutional recruitment and training of civil servants too often remains inflexible, making it hard for employees to move in and out of public administration.
Technology will make possible more porous organisational boundaries.10 The digital world allows skilled workers to come together from around the globe to design and operationalise particular public sector projects. In my recent report to the Australian government, Learning from Failure, I called this the “Hollywood model”.11 It is based upon the concept of a creative workforce bound together temporarily by a shared common interest in completing a particular task, and then dispersing. Technological connectivity will allow people to contribute to beneficial public impact, without having to commit to a lifetime career as a civil servant. Much of public service can become a “flash organisation” assembled for a particular purpose.
The technological future of Public Service 4.0 is already with us. Digitisation and robotic process automation will progressively transform how public administration is undertaken. How fast depends on the scale of government investment and the breadth of its imagination. But increased productivity and improved service delivery are just the starting point. Creative minds, assisted by machines, can envisage new ways of applying connective technologies to government—either to reinvigorate democratic participation or to reinforce authoritarian power. The outcome rests with humans not robots. The question is whether we can be trusted to make the right decisions.
- Klaus Shwab, “The Fourth Industrial Revolution: What It Means, How to Respond”, January 14, 2017, accessed November 24, 2017, https://www.weforum.org/agenda/2016/01/the-fourth-industrial-revolution-what-it-means-and-how-to-respond.
- Shwab, “The Fourth Industrial Revolution”.
- Audrey Lobo-Pulu, “Why We Need an Open Model to Design and Evaluate Public Policy”, January 10, 2017, https://opensource.com/article/17/1/government-open-source-models.
- Richard E. Susskind and Daniel Susskind, The Future of the Professions: How Technology will Transform the Work of Human Experts (Oxford University Press, 2016).
- William D. Eggers, David Schatsky and Peter Viechnicki, “How Artificial Intelligence Could Transform Government”, Deloitte Insights, April 26, 2017, accessed November 24, 2017, https://dupress.deloitte.com/dup-us-en/focus/cognitive-technologies/artificial-intelligence-government-summary.html. See also Gary Flood, “AI to Have Huge Impact on America’s Public Sector, Says Deloitte”, June 1, 2017, accessed November 24, 2017, http://www.thinkdigitalpartners.com/news/2017/06/01/ai-huge-impact-americas-public-sector-says-deloitte/; Michael Chiu, James Manyika and Mehdi Miremadi, McKinsey Quarterly, “Where Machines Could Replace Humans—and Where They Can’t (Yet)”, July 2016, accessed November 24, 2017, https://www.mckinsey.com/business-functions/digital-mckinsey/our-insights/where-machines-could-replace-humans-and-where-they-cant-yet; and Tina Hallett and John Hawksworth, “What Will the ‘Rise of the Robots’ Mean for Public Sector Jobs?”, April 6, 2017, accessed November 24, 2017, http://pwc.blogs.com/publicsectormatters/2017/04/what-will-the-rise-of-the-robots-mean-for-public-sector-jobs. html.
- Alexander Hitchcock, Kate Laycock and Emile Sundorph, “Work in Progress. Towards a Leaner, Smarter, Public-Sector Workforce”, February, 2017, accessed November 24, 2017, https://reform.uk/research/work-progress-towards-leaner-smarter-public-sector-workforce. See also Thomas Colson, “Robots Will Take Over 850,000 Public Service Sector Jobs in Britain by 2030”, October 25, 2016, accessed November 24, 2017, https://www.businessinsider.com.au/automation-to-claim-850000-public-sector-jobs-by-2030-2016-10.
- Quoted in Guy Kirkwood, “Opinion: The UK Government’s Big Robotics Opportunity”, posted by Gary Flood, March 28, 2017, accessed November 24, 2017, http://www.thinkdigitalpartners.com/news/2017/03/28/opinion-uk-governments-big-robotics-opportunity/.
- Stephen Easton, “Paul Shetler: Forget High Tech Fantasies if You Can’t Answer the Phones”, August 2, 2017, accessed November 24, 2017, https://www.themandarin.com.au/category/portfolio/education/.
- US Department of Human Services, “Meet Emma, our Virtual Assistant”, updated May 17, 2017, accessed November 24, 2017, https://www.uscis.gov/emma.
- For instance, blockchain technology could be used to record and verify a variety of public transactions and documents, enhancing mass collaboration and accountability. Anticipated government uses of blockchain technology include identity management, document provenance, supply chain regulation and—most ambitious—automated online voting systems.
- Peter Shergold, “Learning from Failure: Why Large Government Policy Initiatives Have Gone So Badly Wrong in the Past and How the Chances of Success in the Future Can Be Improved”, updated February 5, 2016, accessed November 24, 2017, https://legacy.apsc.gov.au/sites/default/files/learningfromfailure.docx. Also see “Public Service Needs ‘Hollywood Model’ of Work”, Canberra Times, February 5, 2016, accessed November 24, 2017,www. canberratimes.com.au.