Exclusive: Talent Shortage is The Key Pain Point in the AI Industry
Exclusive interview with Timothy Leung: Part 1
Artificial intelligence is one of the emerging fintech trends in Hong Kong. While there are some great minds and AI companies in Hong Kong, they can’t survive without sufficient funding and technical support from government and industry leaders. We had the opportunity to invite Timothy Leung, Executive Director of HKAI Lab, to share his thoughts on how HKAI Lab facilitates the growth of AI ecosystem in Hong Kong. He also identified talent shortage is the key pain point in AI industry which hinders integration of AI and blockchain.
HKAI lab has a six-months accelerator program, which is held twice a year. Can you tell us more about this program?
HKAI LAB is a non-profit making entity and it’s fully funded by Alibaba Group and SenseTime Group. I don't think I need to elaborate more on Alibaba Group. But you may also want to know about SenseTime. SenseTime is a Hong Kong startup founded by a professor of the Chinese University in Hong Kong. HKAI Lab was established in May and we started official operation in the Science Park six months ago.
The goal of HKAI Lab is to promote AI technology in Hong Kong. In order to do that, we decided to do two things. First is running an incubation program or accelerator program and we cooperated with the Hong Kong Science Park. We are initially running the program for six months now, for the second cohort onwards, we extended to 12 months to provide adequate support for the startups. Note that we accept new application every six months but the cohort can stay in the LAB for 12 months.
What our accelerator program excels is that we're focusing on AI. We are well aware of the skill set required for startups, as well as the hardware needed to perform data crunching. It's like doing mining in blockchain. With that we apply our own GPU supercomputer and we let the startups in our program to use it freely. That is something unique that we are offering. We also provide coaching services to the startups, for example, training courses of the Alibaba deep learning platform and the SenseTime development platform. These tools have different strengths, because Alibaba Group is a master in e-commerce, whereas SenseTime excels in image recognition. We have luncheon with our startups and scientists from SenseTime to discuss the pain points that startups are facing. The scientists can provide useful advice in addressing these pain points and this is another unique advantage of our accelerator program.
What are your criteria in picking the startups into the program?
We not only take in applicants from the startup companies, but also personal applications. So basically a university student, or professor or someone's working in a large enterprise can apply to us. If they have a good idea without a company, they can apply to us attaching their testing idea and business plan first. If we think the applicants have good potential, then we can accept them to the program and they can form a company afterwards. In Hong Kong, it only takes a few hours to get a Business Registration (BR) license, and then they are enrolled to our program.
As AI development is still at its early stage in Hong Kong, we don't have enough AI solution providers. We welcome the brains of creative people, either in academics, or even working in enterprises. That’s basically what we do in a nutshell.
Office of HKAI Lab in Science Park
How do you build the ecosystem in AI industry?
The second key initiative we are doing is to build the ecosystem around AI technology. Having a startup is good, yet they need businesses and revenues to survive. We try to build an ecosystem connecting enterprises and AI startups with a “Pain Point and Showcase approach”. Enterprises like real estate developer, big chain retailers, and banks have expressed their interests to explore AI technology towards their pain points, however, they don’t know where to start as they can’t grasp the fundamentals of AI technology. We then try to match their demand with suitable solutions from AI startups, regardless the startup is from our accelerator program or not. The AI startups will showcase their solutions and that’s how enterprises and AI startups work together. Venture capital firms are another player in our ecosystem, which provides funding to promising AI companies.
Can you share with us some collaboration with academics and scientists?
A good question! The way we're working with academics and scientists is that, if our startup has an issue with AI technology, they may work with a professor in that field. We're leveraging the universities to extend our reach to the students from the entrepreneur center and technology transfer center. I directly talked to the professors and hoped to build solutions helping the startups.
I think the unique position for HKAI LAB is we have this close relationship with the startups in the past. The startups also put faith in us to work with academics to solve that particular pain point. We also have good connections with CTO of Alibaba group and professors from different universities serving as our advisors, which can provide adequate technical support to the startups.
What the key challenges in integrating blockchain and AI?
I think the key challenge is talent shortage in Hong Kong. Finding enough talent in AI is difficult. Whether it is a large enterprise or startup, they all scream for AI talents. We don't have enough well-trained professionals in the AI field. This field is so new, that universities don't have degrees for it. We are picking people from computer science, statistics, physics and mathematics fields, as long as these people have the correct mindset working on AI. It is extremely rare to have someone with expertise in both AI and blockchain.
Having said that, there are some advantages of integrating blockchain and AI because you need a lot of data processing in AI. The protection of data source is a common issue in AI. Blockchain may have a way to solve that since it is a distributed ledger, and you can trace transactions on the blockchain. The immutable nature of blockchain ensures data protection. Another scenario is when you build a financial model that people will use the data to formulate an AI solution, which users can trace back the data with blockchain so they can reward the original data owner. In that sense, I can see the beauties of merging the two technologies.
Currently there are many challenges in integrating AI and blockchain. People also talk about smart contract would be a way of merging the two technology. Smart contract is definitely the scenario for blockchain. But whether the contract is “smart” is an issue for AI. How could we merge the two together? I think there's definitely a potential yet the challenges are still there, like talent shortage and lack of successful integration cases.