23 Aug 2017
AI in the Workforce
Jennifer Van Dale heads the Asia employment law practice of global law firm Eversheds. She has a special interest in technology and how it is transforming society. The long-time Hong Kong resident and member of the American Chamber of Commerce Board of Governors discusses how artificial intelligence (AI), automation and robots will revolutionise the way we work.
Is automation coming to the workforce, whether we like it or not?
Automation is already here – we just don’t always think of it that way. For example, automatic teller machines (ATM) automated the job that bank tellers used to do in the 1960s. Some types of automation started in the 1700s, and it has been implemented slowly throughout industrial history. What’s changing now is the automation of more complex tasks and the speed at which it’s happening.
It’s faster and more complex, because in recent years, we’ve made advances in natural language processing (systems that can understand language) and machine learning (systems that can learn from experience). A good example is Siri: if you use it, then you’ll know that it gets better over time. When combined with the low cost of transferring data, these advances create the potential for automation with more complex tasks that, in turn, have a wider application.
AI can give businesses better insights and help employees make decisions that might have otherwise been above their skill level.
What are the benefits for business as a whole?
There are many benefits for business. At one end, it’s basic economics: automation allows manufacturers to make more widgets for less money. Further along the line, it also can lead to individuals performing at a higher level. The bank tellers who are freed up by ATMs can be redeployed as customer service or sales representatives, for example, instead of just counting money. Apart from automation, AI can be particularly beneficial by providing support to humans (rather than replacing them). For example, an AI-supported sales rep can get suggestions based on an analysis of millions of data points that a person couldn’t hope to do in the same amount of time. AI can give businesses better insights and help employees make decisions that might have otherwise been above their skill level.
What might AI “look like” in terms of workforce planning?
One immediate benefit for Human Resources is the ability to streamline mundane tasks such as scheduling, answering onboarding questions, identifying training gaps and helping to address individual training needs. Having these tasks automated or AI-assisted frees up HR leaders to do more strategic and valuable work.
Looking more specifically at two areas – recruitment and future workforce planning – AI and automation are already helping organisations. For example, Expedia uses AI in its recruitment process to identify job descriptions that might be unconsciously biased. Their data shows that job adverts with words like “manage” tend to attract male candidates, while words like “lead” will attract female candidates. Using AI systems to weed out wording that is unintentionally biased helps attract applicants based on skill rather than irrelevant factors.
A word of caution, however: we need to be careful that we don’t over-rely on AI as an arbiter of objectivity. Someone has to write the algorithm and there are numerous instances where AI has in fact perpetuated bias rather than removed it.
When it comes to future workforce planning, AI can help predict business needs, such as where an organisation’s productivity is likely to go down, or if consumer or B2B demand is likely to grow (requiring an increase in production or staffing). By cross-referencing data points, these types of prediction models can also give HR quick access to staff engagement levels, training needs and compliance risks, which HR can then use to plan more strategically.
Should workers fear AI taking over jobs?
This is the question on many minds – it has enormous implications for businesses, individual employees and society as a whole. It’s important to note that for as long as we have been improving processes, we have also been predicting the end of employment. In the 1500s, William Lee was refused a patent for an automated knitting machine for fear that it would deprive young women of their livelihoods. In the 1800s, tailors saw sewing machines as a threat. At the beginning of the Great Depression in the 1930s, the New York Times warned of “the threat of the machine age.” In 2017, I have seen at least five lists of “safe jobs” or analyses of when a particular job is likely to be replaced by robots.
There are two predominant views of where AI will take us: either we are headed for a dystopian future of the Terminator variety, where “Skynet” rules and humans cower and run for their lives, or we are starting a second Renaissance where humans are finally free of work and can turn their minds to ponder the improvement of humanity for all eternity.
What do you think?
My own view is of course somewhere in the middle – and is underpinned by the belief that without active intervention, Parkinson’s law will prevail (that is, work will expand to fill the time available). Just look at other technologies that were supposed to free us from the tedious: e-mail hasn’t reduced office communications or made it more efficient; smartphones haven’t given us more time; and washing machines and vacuum cleaners haven’t released us from housework at home.
I believe it’s undeniable that AI and automation will cause some people to lose their jobs. It’s also undeniable that new jobs will be created. To succeed in the future, we need to ensure that we are training people for the right kinds of jobs and educating them with the right skills. This requires cooperation from all sides, including academia, science, business, government and labour.
Can companies and their employees coexist happily with machines in the workforce?
I think coexistence is inevitable, and we need to see technology as an opportunity and a tool rather than a threat. We need to remember that we as humans will be able to make more informed decisions with AI supported programmes. However, we also need to know why a machine is making this recommendation or decision before we can trust it and really use it. I believe that when we have AI systems that explain themselves and their decisions to us, we will see a marked increase in acceptance and use of machines in the workplace. There is also the matter of convenience: when we get used to Siri and Alexa making our lives easier, and even come to depend on them, I believe we will easily integrate similar systems in the workplace.