By Gene Lin
Amidst the economic consequences caused by the global pandemic, automation has returned to the forefront of discussions. Across various industries, call centres and customer support are some of the sectors most vulnerable to automation, specifically to a branch of artificial intelligence (AI) called natural language processing — the art of making computers understand and communicate with human language.
With the advent of projects such as IBM Watson and Amazon Lex, natural language processing is no longer science fiction. Yet, while experts have made it clear that the AI revolution is virtually inevitable, it remains unclear what the immediate cost-benefit analysis will look like for businesses that adopt these technologies during a recession.
Recession and Automation Come Together
In a joint statement with his colleagues, Marc Muro, a senior fellow and policy director at Brookings Institution’s Metropolitan Policy Programme, said that while it may seem intuitive to assume automation will slow down during a recession due to human labour becoming cheaper, the entire assumption is, in fact, not true.
“Automation happens in bursts, concentrated especially in bad times such as in the wake of economic shocks, when humans become relatively more expensive as firms’ revenues rapidly decline,” said Muro.
In a similar opinion piece, Carl Benedikt Frey, director of the Future of Work programme at the Oxford Martin School of the University of Oxford, said that not only does automation rise with recessions and recede with wars, evidence from the 2008 financial crisis also suggests major recessions can cause repetitive jobs to be automated away for good.
“Weak AI” and “Strong AI”
To understand the powers and limitations of natural language processing AI today, it is important to recognise the difference between “weak AI” and “strong AI”. Weak AI refers to software designed for very specific and narrowly-defined tasks, such as winning a chess game or identifying images. In contrast, strong AI is designed to handle nuanced and broadly-defined tasks that only functioning humans can do, which remains merely a theoretical concept for now.
Today, most AI softwares on the market would be considered weak AI. While these machines are very good at performing certain tasks – both simple and complex – they are nevertheless limited in scope and capabilities.
“There’s no way for a machine to follow a conversation if the customer gets extremely diverse in terms of what they want to talk about, it’s really not possible. It will take a few decades before that changes,” said Miles Wen, CEO of Fano Labs, a Hong Kong-based company that develops automated speech recognition software for commercial use, with a specialisation in mixed languages, including Mandarin and Cantonese.
Cost and Benefit
Despite the obvious technical limitations, natural language processing AI is already making an impact. For example, Fano Labs has developed AI software such as chat-bots, voice-bots and speech-analysis programmes which are currently used by banks, governments, and telecommunication companies in Hong Kong and China. The customers generally see a 20 to 30 per cent cost reduction in their contact centre operation within six months after the system goes into operation, according to Wen.
“If you are talking about call centres specifically, I think 100 per cent of the workers would be replaced by AI because they are answering very regular and repetitive questions,” said Kam Fai Wong, professor of the department of systems engineering and engineering management at the Chinese University of Hong Kong.
Wong added that while the prospect of replacing every worker in a call centre with natural language processing AI is technologically feasible, the initial cost of making such a transition would still be higher than outsourcing the task to human workers. This means that companies need around two to three years to see the returns.
The immediate implications for the AI revolution could be job loss and restructuring of the labour force. Experts generally agree that many workers will need to acquire new skills which are more complex and harder to automate. In other words, human workers need to equip themselves with qualities of a “strong AI”.
“When you look at the previous industrial revolutions, people are always afraid that they will be replaced by technologies. But that never happened because the market always creates new demand,” said Wong. “The environment has changed, and you should change as well. Opportunities are for the prepared, so make yourself digitally prepared”.
Trying to understand enterprise automation?
As firms need to make decisions in a rapidly changing environment and become future-proof, Lynk along with its 780,000+ experts can help. Contact firstname.lastname@example.org