How AI will deliver cost savings and create new jobs

Sign Up

Get the New Statesman's Morning Call email.

Last year, the government launched the AI Sector Deal along with the Centre for Data Ethics and Government Office for AI. Tim Clement-Jones, the chair of the House of Lords’ first AI Select Committee, warned that the country was falling behind South Korea, Canada and Germany in its AI-sector endeavours.

His speech at the AI Expo coincided with the release of a 181-page House of Lords report that called for a more structured framework for AI development and addressed the importance of grasping its economic, ethical and social implications.

Fast-forward to 2019. In May, the UK announced the first members of its new AI Council. Representatives from the private sector, data privacy organisations and academia were among the appointees.“[The Council] will represent the UK’s AI Sector on the international stage, and help us put in place the right skills, data and ethics so we can make the most of AI technologies,” Jeremy Wright, Secretary of State for Digital, Culture, Media and Sport, said in his keynote address at the Viva Tech conference.

So far, the UK has pledged £115m for 16 new AI Centres for Doctoral Training (CDTs), which will fund PhDs for 1,000 students. The AI Council represents another milestone in encouraging tech leadership, educating the general public and promoting ethical and thoughtful AI adoption in both the private and public sectors.

But some of the most concrete examples of the UK’s AI agenda have been less pronounced, particularly in public sector departments. In 2017, the Department for Work and Pensions (DWP), the government’s largest department, completed a highly successful robotic process automation (RPA) project. RPA refers to “software robots” that function over existing IT infrastructure to automate repetitive, rules-driven tasks.

One of the DWP’s functions is to provision welfare and pension services to about 20m people every year. This includes processing pension claims, which prior to RPA implementation, had been completed manually, resulting in a backlog of 30,000 claims.

Shaun Williamson, senior product manager at DWP, estimated that catching up would require a labour force in the thousands working over many thousands of hours. But by using only 12 UiPath software robots, the DWP cleared that entire backlog in just two weeks. “Our experience suggests the return on investment is around 15:1,” Williamson said.

Despite this public sector success story, RPA has arguably achieved greater public recognition outside of the UK. For example, the European Commission released its Coordinated Plan on Artificial Intelligence in March that, “acknowledges the use of robotic process automation [RPA] and the impact it has had on improving public sector processes.” European Parliament members also emphasised AI and robotics’ transformative potential in the health, energy, agriculture and transportation spheres.

Meanwhile, as the UK moves towards Brexit, taking the lead on AI is no doubt a chance to assert its tech acumen as it confirms an EU departure. In his announcement of the AI Council, Wright acknowledged that leaving the EU doesn’t mean the country is “turning inwards”. He added: “We are still committed to being open, outward looking and optimistic about the world beyond us and opportunities provided by new technologies. And we still care about the positive development of these technologies, and making them work for the benefit of humankind.”

So should “these technologies” include RPA? RPA has clearly demonstrated its potential to serve the public by boosting departmental productivity in the DWP. Meanwhile, the body that the UK hopes to distinguish its identity from has already said robotics are an asset in social and economic digital transformation. Additionally, RPA can function with legacy public sector technologies, and is quick and cost-effective to deploy.

Most importantly, RPA is a gateway to AI innovation in both the public and the private sectors. Unlike AI, RPA is limited to rules-based processes that usually involve structured data. This curbs its ability to complete cognitive processes such as predicting outcomes or processing conversational language.

But when RPA is enhanced with AI, robotic software can continuously improve its efficiency and performance. For instance, it could:

  • Process conversational language.
  • Prioritise certain automated tasks over others.
  • Visually interpret user interfaces to interact with software more fluidly.
  • Contextualise unstructured data by understanding document contents.

UiPath, the same vendor that helped the DWP clear its pension claims backlog, is already investing in these capabilities. As the UK stimulates AI development, it must keep all possible paths to AI top of mind. This will be important for educating the general public, but also for leading by example.

According to a recent UiPath survey, 61 per cent of public sector employees in the UK spend 30 per cent of their time each week on repetitive chores that can be automated by RPA. The majority of respondents (78 per cent) also said that they’re not afraid that automation will take their jobs.

Research from Deloitte supports this sentiment. In 2017, it estimated that automation contributed to the loss of 800,000 low-skilled jobs in the UK, but that it created 3.5m higher-skilled, and higher-paying, new jobs. The World Economic Forum paints a similar picture. It predicts that by 2022, AI and robotics will create almost 60m more jobs than they will destroy.

In the UK’s public sector, automation has predominantly been well-received; 74 per cent of government employees said their experiences with RPA have been positive.

Nevertheless, misconceptions about AI, RPA and other transformative technologies remain. Many employees, especially outside of the 18-24 age range, are sceptical about automation’s ability to make them more productive, despite admitting that they spend a significant portion of their time on mundane, administrative tasks.

This likely stems from a general lack of awareness about AI and robotic capabilities, one that the UK’s AI-forward initiatives will hopefully address in the coming years. The onus is also on the private sector and big tech to contribute thought leadership that helps inform the responsible and ethical progression of the technology.

For now, though, all eyes will be on the public sector to facilitate important conversations about new technologies. Just as importantly, the UK government will need to take a pragmatic approach to the internal implementation of AI and robotics. And technically speaking, RPA looks like the most pragmatic approach of all.


For more information, please visit: www.uipath.com