Taipei, Sept. 28 (CNA) Andrew Ng, a British-American AI pioneer, said during a speech at an AI forum in Taipei on Wednesday that AI is a new “general purpose technology,” that, like electricity, has all sorts of potential applications and is creating numerous business opportunities.
Ng, who founded Landing AI and DeepLearning.AI, and is a managing general partner of AI Fund, began his Taiwan trip earlier this week, during which he has shared his views on how AI could change the world.
He was invited to speak at an AI forum held by the Yonglin Charity Foundation on Monday and another held by Taiwania Capital on Wednesday.
Ng highlighted in both forums that AI technologies are general-purpose, meaning they can be applied in many different scenarios, and that many applications have yet to be identified and created.
“There will be fads along the way,” he warned, adding that although some ideas may be good, they will not be long-term sustainable business models.
He cited Lensa, a portrait app where users can upload a photo and get a digital drawing of it, as an example. Ng said it appeared to be a promising product and that revenue levels were high last December, but that profits tailed off quickly.
Sept. 25: AI risky but has potential to improve quality of life: AI experts
There are, however, many long-term sustainable businesses like Uber, Airbnb, and Tinder, Ng said.
But he pointed out that the current value of AI is “very concentrated in the consumer internet sector,” where one piece of AI software can be applied to billions of users to generate massive economic gains.
One hundred engineers could be hired and a billion dollars spent to work on one project that has the potential to reach a billion users, but outside the consumer internet sector, there is unlikely to be a billion potential customers, he said.
However, in certain industries, a US$5 million investment can still have a significant impact.
Ng spoke about one experience he had working with a company that produces agricultural machinery, saying “We figured that if we can work out how tall a weed is by using cameras that have computer vision, we could then chop the weed at the right height, so there would be more food for the farmer to sell and it would also be better for the environment.”
Ng said this is where opportunities lie, in the “application layer” of the AI stack.
According to Ng, the AI stack also consists of a hardware layer, which includes semiconductor companies such as Nvidia and TSMC; cloud infrastructure, such as Google; and a layer of developer tools, such as Open AI.
However, there need to be many more new applications of the technology to generate revenue streams to fund the companies providing the tools and the infrastructure to ensure the economic ecosystem can function, Ng noted.
Ng said he has been working with numerous experts and that their domain knowledge combined with his understanding of AI has led to fruitful business collaborations.
For example, Ng worked with the former Tinder CEO to create Amorai, a company that combines AI with relationship coaching.
Another example was when he worked with Mitsui, a global maritime shipping company, to create the startup Bearing.ai, which provides software that helps ships arrive at their destination on time using “about 10 percent less fuel, saving about half a million dollars per ship and also reducing carbon emissions.”
Ng believes that because AI is a new powerful technology that disrupts the current technology landscape — as all the rules are being rewritten with some old technologies becoming less relevant and new things being invented — it is a good time for people to jump in and be innovative.
In these changing times, Taiwan could have an advantage in fields that it is already strong in, Ng said.
He cited precision manufacturing as one example, saying “If I wanted to build AI for precision manufacturing, I would have no idea how to do this in Silicon Valley, but there are hundreds of companies that do precision manufacturing here.”
Because of this disruption, now is as good a time as any for Taiwan to leverage its semiconductor manufacturing strengths and flourish in the software part of the industry, the AI expert said.