Artificial intelligence (AI) can potentially deliver benefits to local government in these challenging times, says Philipa Bradley of Oracle.
There are three types of AI:
- Super AI – the large scale walking, talking robots – like those featured in The Terminator. Probably the scariest part of AI to us lay folk but there is very little research being done on Super AI and it is unlikely to play a material part in our day-to-day lives in the medium or long term.
- Strong or General AI – which is about building artificial intelligent systems as good as a human – modelling the mind (eg reasoning, representing knowledge, learning, communicating in natural language and combining all these skills to a common goal).
- Weak or Narrow AI – task based or machine learning – just a small piece of what a human can do. Every time you perform a Google search, instruct Alexa or Siri – you are using Weak AI. The Google Adverts that pop up are through AI as are the Social Media curated news feeds and adverts. Or if you have used a retailer’s website and have a query, which you have typed in, you may be responded to by a chatbot (with a name): you may well think you are dealing with a human – but in fact it is a chatbot (this is known as Masking). Often if you have a supplementary query, you will be directed to a human to take over the interaction, but the first port of call is likely to be a chatbot.
AI requires three things to work: models (a sequence of steps or a formula to inform a decision); data (often from multiple sources); and computing power. The more relevant data you collect, the more likely the machine will learn and make better, more informed decisions.
AI will never replace humans to the degree that is sometimes discussed in the media. Currently, AI systems are just not very good at certain things, such as: physical movement; detecting and responding to feelings; or complex facial recognition.
So, in the future, some of the core roles that local government provides, including social workers, teachers, librarians and refuse collectors, will still be required and technology can help make these people facing services more efficient and deliver better patient outcomes. A simple example of this right now would be routing social workers to use the most efficient journey to visit clients taking into account weather data, roadworks, traffic congestion and indeed the client’s own preferences. Another instance might be the use of robotics in social care (for further details see RIKEN’s Robear).
AI is however much better at some things than humans…
- AI is much more objective, it has no inherent bias or pre-conceptions
- It is better at complex gaming than humans, 21 years ago Deep Mind beat Gary Kaparov at Chess and in 2015 it beat the world Go champion very convincingly.
- Most importantly though, it is far superior in the area of finding patterns and trends in data and can deal with high-volume and scale far better than a human.
There are three main categories where AI can potentially add value to local government:
- Automation and efficiencies
- Enhancing the customer experience
- Data insights to drive better decision making
Just a few examples of the benefits this would drive are:
- Contact centre chat bots helping the public to pay parking fines, understand council tax band rules, or apply for a business licence.
- Touchless transactions, eg automatic requisitioning (replacing bins, ordering new school books); automation of routine or manual tasks (graffiti cleaning); automatic reconciliation and closure of financial ledgers; and intelligent checking of invoices.
- Automation of expenses.
- Better management of the supplier base (search for best prices, best terms).
- Transformed recruitment (sifting hundreds of CVs overnight and presenting best fit ranking for vacant roles to the HR Manager first thing in the morning).
- Risk and compliance proactive alerts.
- Connected intelligence – contact centre automatically requisitioning a third party preferred supplier to remove graffiti; provisioning for cost within the budget; automated payment of invoices to specified terms.
In many areas of the UK and indeed within local government, there are examples of these technologies in use already but the basics need to be in place for true optimisation. Elastic computing, where machine power can be spun up or spun down quickly is really important for these hungry models to do their calculations and searching cost-effectively. Equally using Cloud platforms rather than traditional on-premise systems is a pre-requisite in order to get easy access to the third party data). In short, AI cannot be truly optimised unless councils are using a Cloud platform.
Philippa Bradely is Account Executive (LG & Health), Public Sector Applications at Oracle