I recently discussed artificial intelligence (AI) during an inspiration session I hosted with local government colleagues, and the topic certainly makes for an interesting discussion. Britain is base to some of the world’s most innovative AI companies and is also host to a growing ecosystem of investors, employers, developers and clients. AI is a fast-growing area of development and in February 2017 it was identified as one of the eight areas of research that could be supported through the Government’s new Industrial Strategy Challenge Fund. Keep an eye on social media and an ear on Radio 4 and you can’t help but notice that references to AI are cropping up more and more regularly these days. A really interesting facet of this next phase of the digital revolution is the impact it will have on white-collar workers. In contrast to the first industrial revolution which saw the low educated, working-class being disrupted by the introduction of steam-powered machinery, AI looks set to impact heavily on 21st-century professionals, such as lawyers and doctors through the introduction of algorithmic technology. It is therefore perhaps obvious that this next industrial revolution will be equally, if not more disruptive than the first.
There are already examples of how AI is starting to disrupt the financial sector. In March 2017 one of the world’s largest investment houses, BlackRock, announced ambitious plans to use algorithms to manage some of its active equity funds, which focus on quantitative and other strategies that adopt a more rules-based approach to investing. But there is a growing amount of evidence which suggests that other sectors will also be disrupted. The Guardian published an article in February 2017 which quoted a report written by the thinktank Reform, suggesting that around 250,000 public sector workers could lose their jobs to robots over the next 15 years, with AI ‘chatbots’ replacing up to 90% of Whitehall’s administrators by 2030. One local authority in London has already started down the route of using AI to deliver some of their services. Enfield Council worked with IPsoft, an American artificial intelligence company, to build ‘Amelia’. Amelia will work alongside existing front line services to answer resident queries and authenticate licenses and permit applications. Amelia will use natural language processing to learn how to interpret the emotion expressed by residents during their conversations and will learn to respond appropriately. The hope for the council is that callers won’t even notice that they’re not dealing with a human. The justification for introducing this technology echos that of the last industrial revolution; the council seeks to reduce tedious labour and free up employees to do more complex and enjoyable tasks and focus on more subtle forms of dialogue, judgement and interaction. This time, the focus is currently on moving people onto more fulfilling jobs, rather than replacing them, but I guess we will need to see how that pans out as this type of technology becomes ubiquitous.
Whilst Amelia has a voice, perhaps one of the most popular applications for the automation of public services will come in the form of chatbots. In a programme broadcast on BBC Radio 4 back in April 2016, Rory Cellan-Jones, the BBC’s Technology correspondent, and Dr Sabine Hauert, Lecturer in Robotics at Bristol University, discussed the decline of apps in favour of chatbots, allowing users to type instructions, questions and orders – and receive intelligent responses in return. Instead of opening an app-based form to pay your council tax, you’ll simply need to text the bot and request ‘I want to pay my council tax’. These types of bot are already starting to impact on government, even if they aren’t always developed from within. A British born Stanford University student, Joshua Browder, recently designed a chatbot that overturned 160,000 parking fines and also helped vulnerable people apply for emergency housing. The homelessness bot has had more than 3,000 users, with more than 240,000 messages sent and received. Joshua has now updated his robot lawyer, DoNotPay, to give free legal aid to refugees seeking asylum in the US and Canada, and asylum support in the UK.
Socially conscious bots are receiving a great deal of thought and attention from some of the world’s leading experts in AI. Demis Hassabis, neuroscientist, computer games designer, entrepreneur, and one-time world-class chess player decided to give up the professional chess game at the age of 11 because he thought that great minds should be put to better use, helping to cure cancer or find solutions to other problems facing humanity. As an adult, he co-founded DeepMind with the aim of solving intelligence and then using intelligence to solve everything else. Indeed AI is now helping to if not cure cancer, recognise it. In January 2017 a wired magazine report introduced an AI technology which can spot skin cancer as accurately as a doctor. The AI was trained on an image database of 129,000 images and can now perform as well as trained medical professionals. The algorithm developed by Stanford University was fed 129,450 training images of 2,032 different skin diseases (broken into 735 groups to account for diseases with a small number of images). After learning what cancers typically look like, the software was tested against the clinicians in three areas: keratinocyte carcinoma classification, melanoma classification and melanoma classification using dermoscopy. As the AI continues to learn it will undoubtedly get quicker, and perhaps even better at recognising cancer than its human counterparts. The challenge for the technology developers is to now turn what they have created into a diagnostic smartphone app and apply the deep learning approach to other applications which could benefit from the ability to quickly identify patterns in images.
There seem to be no limits to how AI technology can permeate our lives. A London based startup (which I have to say puts me very much in mind of the Black Mirror “Greta cookie”) wants to make controlling a smart-home more natural and intuitive. In a Forbes article entitled ‘The House That Learns: Bringing Artificial Intelligence Into The Home’ the co-founder of the startup, Dagham Cam, describes how his AI technology will incorporate six cameras to learn where objects are in the room, recognise certain people and to respond to a range of motions and gestures – moving beyond the voice command or apps used by personal assistants such as Amazon Alexa. The aim is to make home automation as easy as asking a friend to turn on a light, by simply asking and pointing at the light you want to activate. The technology will utilise reinforcement learning, meaning that over time it should be able to pick up the natural gestures and voice idiosyncrasies of frequent users. It seems obvious to me that this, or similar technology could also incorporate social welfare functionality in order to help combat loneliness or spot when people have fallen, even spotting and advising on potential fall hazards.
The think tank, Reform, suggest that as well as 90% of Whitehall administrators, bots could also replace tens of thousands in the NHS and GPs’ surgeries by 2030, saving the government around £4bn a year. It is therefore unsurprising that the UK government is taking notice of the technology and investing heavily. A press release published by the Department for Culture, Media & Sport and The Rt Hon Karen Bradley on gov.uk on 26th February, announced a £17 million boost for the UK’s booming artificial intelligence sector, suggesting that the AI sector could add an additional £654 billion to the UK economy. So it’s certainly going to be a case of when and how, rather than if, our lives are going to affected (positively or negatively) by the technology.
These are exciting times for those of us working in the public and social sectors, affording us the opportunity to start thinking about ways in which we can exploit bot technology for the benefit of our citizens. It looks like the robot age is here, not perhaps as we all thought – humanoid assistants are probably some time off, but faceless algorithmic-based bots are on the rise. How will this affect us humans? We will have to wait and see. As the digital revolution moves forward, I think perhaps a challenge for governments is to make sure that they design policy to allow full utilisation of the technology as it becomes available, ensuring that public sector AI doesn’t just end up becoming ‘automated telephony 2.0’.
Simon Penny 