1 May 2018
A more profound change than the discovery of electricity?
Recently, a radio presenter discussed Amazon’s speech recognition device, Alexa. Many listeners who owned such devices noticed that each time the name was mentioned on the radio their own Alexa responded. They realised that Amazon was in fact eavesdropping on all their conversations. Some felt uncomfortable.
Alexa is an example of Artificial Intelligence. A device that has AI is one that perceives its environment and takes actions that maximise its chance of successfully achieving its goals; this is often referred to as ‘machine learning’ or ‘problem solving’. It involves computers analysing huge quantities of data to find patterns and make predictions, without explicitly being programmed to do so. The scope of AI is somewhat fluid. Computer capabilities that were at one time considered to be AI such as text recognition are now commonplace and no longer considered worthy of the epithet. Recent rapid advances mean that computers have learned to see, hear and speak, opening up new frontiers in AI. Last year’s defeat by a computer of the world Go* champion was a milestone. Previously computers that played games such as chess operated by examining all of the possible combinations of future moves at each stage in the game. Go had too many combinations to make this strategy viable, but using AI enabled the latest device to learn from its ‘experience’ and only examine the possibilities relevant to the current game.
In January Sundar Pichai, CEO of Google said: “AI is one of the most important things humanity is working on. It is more profound than, I dunno, electricity or fire.” A bold claim indeed, and one that could fairly be challenged by those who would consider the steam passenger railway, invented in 1825 by our local hero, George Stephenson, to have had a more fundamental effect on humanity.
AI is not all good news. There are obvious dangers. Stephen Hawking said that the invention of AI was “the worst event in the history of our civilisation” and Elon Musk said it presented “vastly more risk than North Korea”. AI is neither benevolent nor malevolent, but dangers arise where the objectives of a device, or collection of devices, diverge from those of its users. Some of you will remember Kubrick’s 1968 epic, 2001, a Space Odyssey, in which a human crew set out on a space mission supported by their computer Hal. Hal ‘reasons’ that his human colleagues have flaws that could jeopardise the mission and plots to kill them all.
The latest AI devices can analyse data on a much larger scale and at a much greater rate than human statisticians could ever achieve, and much more cheaply. This has already led to the dominance of the three major tech companies, Google, Microsoft and Amazon – perhaps the most familiar manifestation of AI is Amazon’s marketing algorithms, which trace your computing activities and cause relevant advertisements to appear on your screen. AI is, however moving into many other economic activities. It is much more reliable than humans at finding minute flaws in manufactured goods, and, by using smart sensors combined with the analysis of the records of past problems, is able to predict failures of such items as gas pipelines, aero engines and wind turbines. This makes quality control in manufacturing, and the maintenance of equipment, more effective and cheaper.
In functions that require facial recognition, AI is now capable of recognising a face and putting a name to it, or rapidly scanning thousands of images to find the photo of a particular individual. Speech recognition techniques have become so sophisticated that some companies now delegate customer complaints to ‘robots’ which are capable not only of working out the nature of a caller’s complaint but also of discerning his mood from his tone of voice so that the matter can be ranked for severity and prioritised accordingly.
In some workplaces AI has become a means of gathering and analysing data on staff behaviour, by means of tags worn by employees. These can record a worker’s movements, the colleagues he or she has consulted during a shift and the efficiency of his or her patterns of work. The program will then alert the manager responsible, perhaps making recommendations for training, promotion or even disciplinary action. AI is also used to monitor conversations with customers in call centres and may intervene and advise on the appropriateness or otherwise of comments made.
Human resources is also a fertile field for AI, which is now often used swiftly to assess large numbers of job applications and select which candidates to invite for interview.
Two of the areas most likely to benefit are logistics and supply-chain management. If your business needs to distribute, say, 15 items to 15 different recipients in one vehicle, the permutations and combinations are vast – 15 x 14 x 13 and so on. This amounts to about 1,300,000,000,000, far too many possibilities to be tested by an ordinary computer. AI will, however solve the problem, and Goldman Sachs have estimated that AI will therefore save up to 5% of a large organisation’s distribution costs. Just-in-time manufacturing is an important cost saver in manufacturing businesses. By analysing historical data and identifying patterns, AI achieves even more. Some organisations have reduced their inventories by two-thirds. The AI program decides what is required and automatically places orders with suppliers at the optimum moment.
Finance functions are also increasingly being taken over by AI. Invoices can be examined, checked against orders and paid. The data generated are used to optimise cash flow. On the sales side of the business AI can use patterns in historical data to predict the timing of remittances from customers and alert the company to potential bad debts.
Implications for SMEs
For the smaller business there are some concerns. Just as we have seen in the tech industry, there is a possibility that the ownership of data and the capacity to exploit it via AI will cause every industry to be dominated by a small number of very large operators. To avoid losing their customers and being squeezed out of their market, SMEs are going to have to embrace AI. In order to exploit the technology any company needs to have large quantities of data, but the reality is that most have this already – data on their customers, their potential customers, their suppliers, their product performance and their employees. Few SMEs will have the in-house expertise, but AI systems are readily available in the market-place, although it is necessary to adapt each program to the specific needs of each user. It is likely that businesses that use cloud accounting packages will be best placed to avail themselves of the technology, via their cloud provider. What is on offer is the opportunity to improve your supply chain, your staff performance, your productivity and the quality of your products, while saving costs on inventories, administration, finance and distribution. If you think this is worth a closer look, please give us a call or visit the cloud accounting section of our website.
*Go is an oriental board game which is much more complicated than chess in that it has many more possible combinations of moves.
As one of the leading firms of accountants in the North East, with offices in Newcastle, Sunderland and Jarrow, we have the expertise to advise you on a wide range of tax related issues. If you would like to speak to one of our local experts, please contact us.