Chatbots, AI news readers: Asia embraces automation despite job loss fears

Indian entrepreneur Suumit Shah took to Twitter last month to describe his experience with generative artificial intelligence: "We had to layoff 90 per cent of our support team because of this AI chatbot. Tough? Yes. Necessary? Absolutely."

With the chatbot, queries to his e-commerce site dukaan.com took just over three minutes to resolve from more than two hours earlier, and customer support costs fell by about 85 per cent, said Shah, the chief executive.

Shah's tweet, which has had more than 2.6 million views, drew mixed responses, with some congratulating him for embracing the technology and getting good results, while others berated him as "heartless" for laying off nearly 30 employees.

"It's not me who is going to decide the fate of hundreds of thousands of people employed in areas where generative AI is being deployed; it's up to them to see how AI is changing their jobs and learn, as not all employers have the resources to train employees," he told the Thomson Reuters Foundation.

Generative AI has been widely hailed as the next big growth driver in technology. The launch of San Francisco-based OpenAI's ChatGPT chatbot to the public in November set off a global frenzy, with more than one million downloads in the first week.

Since then, the AI-based tool and image-generating tools such as Midjourney have been incorporated in education, marketing, customer service, online search and content creation, with users touting greater efficiency and reduced costs.

But lawmakers and some tech experts in Asia have warned of a lack of regulation to prevent misuse, the risk of privacy violations and misinformation, and the potential for income inequality as certain types of jobs are automated faster than others.

"Unequal access to AI tools can worsen socioeconomic disparities, favouring those with greater resources and leaving marginalised communities behind," said Kazim Rizvi, founding director of The Dialogue, a policy think tank in Delhi.

"This imbalance can lead to a concentration of power and influence in the hands of a few, while others are left behind without the means to participate fully in the AI-driven economy," he said.

Asian nations have been slow to regulate AI or introduce ethical frameworks, lagging the US AI Bill of Rights that offers guidelines for the responsible design and use of AI, and the EU's proposed AI Act that imposes tough standards.

Skewed datasets

About 18 per cent of work globally could be automated by AI, and 300 million full-time jobs could be lost to generative AI, according to Goldman Sachs, which estimates that AI could eventually increase annual global GDP by seven per cent.

While fewer jobs in emerging markets are exposed to automation than developed markets, more than 10 per cent of jobs in India and nearly a fifth in the Philippines could be lost to AI, Goldman Sachs said in the report earlier this year.

In China, AI-generated models are fast replacing human models in fashion catalogues and on shopping websites. An AI-generated model even featured on the cover of Vogue Singapore's March issue, pitched as "fashion meets the AI revolution".

China also launched what it claimed was the world's first AI news anchor in 2018. Since then, they have also been introduced in Kuwait, Malaysia, Indonesia and India, where networks claim greater speed and cost savings.

In India, AI anchors — from Sana in a skirt on Aaj Tak channel to saree-clad Lisa on Odisha TV — present news bulletins, horoscopes, weather and sports updates in English, Hindi and several regional languages.

Also in India, the high court for Punjab and Haryana states earlier this year turned to ChatGPT to decide whether to award bail for a suspect in a murder case "to present a broader picture on bail jurisprudence where cruelty is a factor".

But while technology is aimed at efficiency, "efficiency does not always amount to justice," said Urvashi Aneja, founding director at Digital Futures Lab, a research collective.

"Skewed datasets used to build these databases for AI tools can lead to biased outcomes, especially for marginalised groups who may be over or under-represented in datasets," she said, adding that there are also risks of exclusion and privacy violations.

"We were hoping that AI will help reduce our workload, but what is actually happening is it seems to replace human labour."