AI trust issues outweighed by potential: Panel
Daphne Yow
ARTIFICIAL intelligence’s (AI) flaws may be many, but the benefits outweigh them, said an Alibaba panel discussion on how AI will impact the global business landscape.
One big area of concern is trust, as large language models (LLMs) are trained on a huge amount of public data which may contain errors or have biases. This results in occasional errors or private information, said Dr Chua Tat-Seng, professor at NUS’ School of Computing.
“Most of us have heard that they tend to hallucinate, meaning that they make up answers to questions that do not exist, and that can be very harmful,” he said.
However, to information-based businesses, AI will remain indispensable. LLMs are good in analysing large amounts of data and correlating with external information, added Dr Chua.
In addition, AI can greatly enhance productivity for simpler tasks in organisations, said Mohan Jayaraman, expert partner at Bain & Company.
Companies can now access generative AI via application programme interfaces (APIs), which allow for data to be exchanged almost directly with the AI without having to go through another software or middleman. The APIs provided by specialised AI companies allow organisations to leverage the technology to grow their businesses, added Jayaraman.
BT in your inbox
Start and end each day with the latest news stories and analyses delivered straight to your inbox.
The other panellists are Zhang Kaifu, vice-president of Alibaba International, and Tatiana Palhares, chief product officer at South Asian e-commerce platform Daraz.
There is a palpable fear about AI taking over jobs and rendering workers obsolete, but AI is no miracle worker.
While AI has the potential to change how people work, Jayaraman noted that to get them accustomed to AI and be able to use the technology correctly require much work. Early movers among enterprises in this space could see a huge advantage as well.
SEE ALSO
“One of the things that I point out about generative AI is that we’re on top of the hype cycle, with the assumption that AI can come in and solve all problems,” he added.
There are still many tasks which AI will struggle with, notably with complex systems and logic, such as getting human relationships right, said Dr Chua.
“It’s a question of whether we want to see this as a glass half-full or glass half-empty (situation),” said Jayaraman.
Copyright SPH Media. All rights reserved.
