AI may not be a silver bullet, but it is definitely here to stay, according to panel members

On the first day of Echelon Asia Summit 2017, KeyReply Head of AI Research Lim Swee Kiat opened the panel discussion on artificial intelligence (AI) with a question: Is AI nothing more than just a fashion fad, or is it really a revolutionary game-changer?

“Objectively, there has been spring and winter in the past decades. But I think we are living at a really interesting time right now, at the intersection of four sectors: infrastructure, algorithm, building on past decade of research, data from internet and storage,” he explained.

“There is a huge potential and AI will continue to become a revenue generator for companies,” he stressed.

Panel members SWAT CTO and Co-Founder Jarrold Ong and Cocoon Capital Managing Partner William Klipgen agreed that AI is more than just a one-time hype; but SmartCow CEO Annabelle Kwok has a different opinion.

“AI is definitely a lot of hype, but then again a lot of hype means a lot of hope,” she said.

She then proceeded to telling a story about how a client expected her team to “copy the entire Amazon Go for me.”

“The hype has pushed it so far that people has a very conceptual idea of what AI can do. But like any technology, AI has its limitations,” Kwok said.

“To some extends it’s very important to communicate [the extent of AI capabilities],” she added.

Does this hype create its own problem for industry players?

“[People are asking] why don’t we use AI for this, for that? That’s where you can get really frustrated trying to explain to them. Even if you’re using it yourself, you’re still not the expert, [as] AI is always changing everyday,” Ong said.

Also Read: US may restrict China’s AI investment, citing security fears

“The problem that I am facing as an investor is deciding in a single startup: what is hype and not hype?” Klipgen added.

In dealing with this situation, it is important to find a balance between the hype and the hope, Kwok said.

“The reason [two AI winters] happened previously was because there was so much hype, everyone is throwing money for this technology. When money is coming in, startups have this tendency to overpromise … while the technology is not as advance as they look to make,” she stressed.

“Because of these overpromises, they did not deliver their promises, so investors are pulling their funds,” she concluded.

The situation in Southeast Asia

 

While the Southeast Asian (SEA) region has begun to see the rise of startups working in the field of AI, Klipgen is cautious about dubbing Singapore as the next hub for AI innovation, though he admitted to have seen a small number of startups with “great quality” of works.

“The center for AI innovation happens still in Silicon Valley. Singapore is traditionally a place where you see less cutting edge innovation and more app-based companies that try to access the local market; that’s basically what we’ve been investing in the last 10 years,” he said.

Also Read: Now In Shenzhen: This Chinese company is building an artificial intelligence robot using VR

He also explained how AI companies in Singapore need to work extra hard to “be cutting edge” and compete in a global scale.

“It requires so much more in terms of resources and being cutting edge, and the competence of the founders themselves,” he said.

So how can companies create an edge for themselves, especially when faced with competitions from tech giants such as Facebook and Google?

“Companies need to explain what they are doing better and probably speak more about how they are similar to, or how they differentiate from the global players, so that we get a sense of where they are in terms of global competition,” Klipgen answered.

Especially that when it comes to the challenges that AI startups in SEA are facing, many of them had to do with dealing with clients and their expectations.

Ong stressed the importance of double-checking for safety, and not letting the programme “doing something that it is not expected to.”

While Kwok felt that the problem lies in the fact that clients “are not willing to pay.”

“When you have AI deployment, you need good cameras, you can’t be using your own camera system. You need to install a server. Many companies said “Yes, I want AI” but when you tell them the upfront hardware cost that they need to do, to integrate inside their premise, this gives them a scare,” she explained.

“‘Show me other instances where AI works [without having to use such expensive hardwares]!’ But as AI is quite new, there are not many use cases that I can show them,” she added.

The post Live from Echelon, AI startups need to communicate the extent of their tech appeared first on e27.