CIO Folder: The jobs threat of automation
14 March 2019 | 0
We all know the ramifications of Artificial Intelligence (A) — or think we do. Last year in this column, we posited that AI was still at an automation level, lacking the human factors of intuition and imagination. That is true at the present and for the foreseeable future. AI is advanced automation dealing with online purchases and accounts, financial dealings like credit ratings and other business processes based on IT. The other path of development is autonomous vehicles based on AI, nowadays rapidly progressing on land, sea and air. Drones are the headline temporarily, driven by hobbyists although military use is far advanced by comparison.
Machine learning (ML) as the basis for automation is a major line of progress. Industrial robots are in the vehicle assembly industry since the early Sixties, developing as complementary IT and programming for repeat processes developed. Today ‘robot arms’ are the mainstay of assembly and spray painting in all vehicle assembly plants. Human labour — however expert — is a lesser partner to automation.
“The real deal is the almost infinite universe of automation. The potential spans autonomous vehicles (land, sea, air), healthcare, industrial or business processes including the most sophisticated financial ones, and state agencies and government. AI is the next generation of powerful ICT automation, a surge of progress”
The advance of AI continues with some pace, accelerating in fact. Leaving aside the science fiction aspect of human-like robots, AI and machine learning are contributing more and more to business, state agencies and research. They have the potential to contribute to every field of human activity. In many respects, their ‘intelligent’ performance may outstrip humans — faster complex calculations, for example, and faster actions/reactions as a result. AI may very well prove to be superior to human intelligence in many areas with an ICT base.
There is a study published last month from University College, Cork which was a combination of social and IT research. “Automation in Irish Towns: Who’s Most at Risk?” stirred headlines. Yet in the opinion of CIO Folder it was superficial, applying an ‘automation risk methodology’ to the 2016 Census. Its conclusion was “The impact of automation in Ireland is going to be felt far and wide, with two out of every five jobs at high risk of automation.” There is list of 10 towns at most risk of automation damaging employment and of 10 towns least at risk.
The towns listed are headed by Edgeworthstown, Ballyjamesduff and Cahir most at risk from automation. The jobs most at risk of automation include office, secretarial and administrative support positions, process plant operators and jobs in agriculture and customer service. We are all in ICT aware of automation looming in administrative processes, but in specific industries e.g. finance and insurance, state services, etc. SMEs throughout Ireland’s small towns are unlikely to be to the forefront of automation, apart from smarter ERP and other IT systems. As for agriculture, the field is so wide that the prospect of automation is confined to specific tasks — on a large scale. ‘Customer service’ physically is not in danger in retailing or hospitality. Online there is a big prospect of automated processes. It depends where it is located but Ireland’s small towns are unlikely.
The study finds that the “likelihood of jobs in towns being automated is explained by population differences, by education levels, age demographics, the proportion of creative occupations in the town, town size and differences in the types of industries across towns.” This strikes as a contrived socio-economic study with dubious conclusions. Education and demographic levels in Irish small towns are dominated by third level which means students migrate to the cities. Besides which the named towns are too small to be demographically or statistically significant. It is the cities versus rural towns that is meaningful — by and large the nation as a whole.
‘Two out of every five jobs’ at risk of automation sounds valid. It is somewhat of a guesstimate. There is no doubt that business processes will be automated, by machine learning and AI. Will the economy expand to offer a balance of jobs? The construction sector is an example. On-site skills will always be there e.g. crane operators. The impact of IT (and AI) in the office processes is growing, such as quantity surveying and estimating, procurement and supply chain. But the real human labour is complementary and essential, starting with the supervisory judgement in spotting difficulties on the site.
Factories will be increasingly automated. That has been true for decades. Mass manufacturing plants have developed robotic processes with increasing sophistication while processing plants, from pharmaceuticals to drinks to food, nowadays have a characteristic of a small human team 24-hour supervising while volumes grow. Mechanical processes have matured to IT-aided actions and in due course we can expect AI to be a platform.
The point is that the economy is growing and the effect on employment and the total of jobs could be positive. The culture of business is already permeated by AI and automated processes. In Foreign Direct Investment and the technology multinationals already here, the norm is sophisticated IT and automated business processes.
It is today’s pervasive culture. AI has the potential to contribute to every field of human activity, beyond manufacturing and business. In many respects, their ‘intelligent’ performance is superior to human judgement and decision-making, particularly in speed. Modern systems enable faster complex calculations, for example, and faster actions/reactions as a result. Every aspect of human activity will be supplemented and enhanced by AI.
For instance, AI and analytics are already a deep and powerful combination. Analytics is an applied science, with more than a touch of art. Knowledge and experience of the domain, the sector or line of business, is essential to the value and relevance of analytics. AI is the technology that will complement and augment that human knowledge and judgement. It is the key to rapid progress in business, science and research, healthcare, education, public administration and other useful fields. It will also contribute to entertainment — from video to gambling — and domestic and personal applications including house and appliance management and disability assistance across a wide range.
The progression of AI and ML will enable two major lines of advance and development: faster computing and automated processes. Sounds simple. But AI will enhance the value of our engineering development in contributing the advances in chip power, memory modules and miniaturisation. As always, what was high-end computing in a previous generation becomes commonplace. That will increase the speeds of smart networks, from telecommunications and the Internet to data centre links and on-premises networking.
But the real deal is the almost infinite universe of automation. The potential spans autonomous vehicles (land, sea, air), healthcare, industrial or business processes including the most sophisticated financial ones, and state agencies and government. AI is the next generation of powerful ICT automation, a surge of progress.
But we have to be wary. Everyone knows that invisible or closed systems have their perils. AI Automation will compound the threats. The more powerful the technology the more dangerous as the counter-technology will be equally capable. Ironically, AI is likely to be our strongest protection in a world of cyber danger.