“The challenges are more around people and processes, and not so much about the technology itself. And the reason for that is there’s been a lot of fear and uncertainty because there have been sweeping statements from some US customers who say, this is great, I can replace huge swathes of people with this technology,” Yap said in response to a question from ET.
She was speaking at the Let’s Talk AI event by Google Cloud at the company’s Asia-Pacific headquarter in Singapore.
What business leaders want is to set up safe AI for their employees, which augments people’s jobs, enables deep analysis with more accurate insights, while replacing certain mundane tasks, Yap said. At the same time, for regulated industries, she said it is important to understand where the role of the tool ends and where humans must play their part.
Further, customers must think about leveraging AI from a growth-first perspective rather than a focus on efficiency, Yap said. “If you only look at the technology and try to apply it to efficiencies, you miss what the future can actually be for your business.”
On returns on investment for generative AI, Yap said that the value businesses get from AI is always going to be greater than its cost.
Discover the stories of your interest
“I think we are charging less than five cents per (generated) image. But the value – the time saved, the time to market – those are the intangibles that Google cannot measure as a business value benefit but our clients and customers can measure,” she said.Typically, most businesses look for an at least 10% uplift through AI adoption, Yap said, adding that regulated industries have higher budgets for AI.
Google Cloud has been working with C-level leaders to understand their business value journeys and help map out how to apply the technology, she said, adding that customers are under pressure to lay out their AI strategies for board members now, especially amid tightening costs, a slowing global economy and uncertainty in an election year.
To reduce compliance burdens and cost inefficiencies, Google is in talks with governments to encourage interoperability of AI regulations across markets, or mutual recognition programmes where certification of an AI model in one country as being aligned with an international standard will be accepted in other countries that also subscribes to that standard, said Chester Chua, Google’s AI policy lead for the Asia-Pacific region, speaking at the same event.
To boost AI adoption across industries, Google as an AI model developer, wants to take the burden of compliances off customers as much as possible, by, for instance, indemnifying them in case of potential copyright infringement suits, he said.
For policymakers, the main concerns are around data, misinformation and bias, safety and security, and risk management, Chua noted.
The reporter was in Singapore at the invitation of Google.
