ARTIFICIAL intelligence is an oft-used phrase in modern parlance. But how do you define this hot buzz word and what it means to you?
That was the lead-off query at the Agoda Outside conference on artificial intelligence (AI) and machine learning.
John Brown, the new CEO of online travel-booking platform, Agoda, headquartered in Singapore, asked his colleagues for a definition of AI, and nothing definitive emerged.
That’s the nature of the evolving beast that promises to be mankind’s future — be it boon or bane.
“AI collects a lot of data,” noted Brown. Not all of it is used. “But what we don’t use today, we make use of another day.”
Varon Zeidman, CTO of Agoda, commented that “without using AI technology today, it’s going to be hard to become a great company”.
Among the panelists at the session, Dileep Rangan of IBM Watson felt AI would have to be “consumable” before it would have a stronger impact on the public.
“It has to go from expert AI services to citizen AI services.
You Might Also Like To Read:
Dilip added that companies should dream big but start small. This allows them to have a big picture but tackle issues in more manageable bite sizes. It helps to boost confidence and doesn’t become an overwhelming challenge.
More Or Fewer Solutions?
Idan Zalzberg, VP Data Engineering at Agoda added that companies need the data to get a broad definition, which would allow them to get quick solutions.
Grab’s head of data science, Lye Kong-wei recalled how machine learning impacted real life with their drivers.
Initially, the drivers were happy to get a long list of possible fares to pick up. But as the machine learning kicked in and the drivers’ driving patterns were better charted, the system picked more relevant jobs for the drivers.
While Grab thought this was more productive, the drivers didn’t like it. “We are getting fewer jobs, they complained,” recalled Lye.
In the end, the programme was shelved.
The importance of trust and relevance are important when AI comes to play, explained Zalzberg: “Consumers are looking for ‘explainable AI’. For example, I may show you fewer options when you search for something, but the goal is not to hide anything from you. As a consumer, you need to trust that I’m showing you what’s relevant to you. And the best way I can show this is by explaining why – hence, explainable AI.
“As we continue to experiment in this space, we need to remember that it’s all about establishing a conversation between the user and the AI to create a much better experience overall,” Zalzberg added.