Chai Wutiwiwatchai is the first director of the National Electronics and Computer Technology Center (Nectec) to be promoted not from a deputy director position, but from the top job at a research unit.
Nectec has initiated many advanced research projects related to computers and technology, most of them off-the-shelf products, but they have not been widely known publicly.
As the latest top executive, Mr Chai says he'll push technology to serve massive user groups in order to make an impact on the country as a whole.
Artificial intelligence (AI), chatbots, machine learning, big data and many other technologies that are emerging as focal points today have been on the radar of Nectec's research teams for a decade. The human-computer communication research unit under Mr Chai's supervision worked on such technologies for a long time.
Mr Chai started working at Nectec in 1995 as an assistant researcher in the speech processing section, following a suggestion of his adviser at Chulalongkorn University, where he earned a master's degree in digital signal processing.
Back then, speech-to-text was unpopular and there were few practical machines for the technology, but Mr Chai enjoyed working on the AI. It was his first exposure to neural networks, and he's been fascinated with the field ever since.
Mr Chai discovered that many countries had done research on speech processing, and he became one of the key people at Nectec who laid the foundation of speech-to-text technology. Having worked for three years, Mr Chai got a scholarship to study for a PhD in Japan.
He was inspired by Sadaoki Furui of the Tokyo Institute of Technology after reading the professor's books, so he contacted him and was accepted as an advisee.
"He is very gentle, a Japanese professor who speaks English fluently," Mr Chai says. The adviser suggested to him that the speech recognition technology he hoped to study was outdated in Japan. Mr Chai recalls his shock that a technology almost unknown in Thailand had been developed in Japan for 15 years.
The professor recommended studying the spoken dialogue system, now commonly referred to as chatbots. This was nearly a decade ago.
In 2010, Mr Chai, as head of Nectec's human language technology lab working on speech-to-text and text-to-speech, tasked his team with developing an AI-powered chatbot. This first chatbot was called Abdul (for Artificial Buddy U Love) and ran on MSN. Back then, companies dismissed the product, likening it to a toy. Years of searching for a market turned up no interest. Now, nine years later, everyone is talking about chatbots.
Nectec, founded in 1986, is considered a pioneer in proof of concept. The key issue, however, is how to harvest the lessons learned and make practical use of new technologies. These are the major challenges facing Mr Chai as director.
"AI back then was not intelligent as today," he says. "It needed a great deal of data. We had the data, but we lacked the processing power. The technology needed to wait until we had better algorithms, and now [graphics processing units] can process powerfully, and users would like to see AI, as most people today have a smartphone. Those are the three critical factors for AI."
Several research projects of Nectec, like the text-to-speech synthesis program Vaja, the speech recognition engine Party and the Abdul chatbot, are off-the-shelf products now and can be used to support various demands of the market.
Nectec has done natural language processing for many years and has transferred technology solutions when companies have asked for them. But, due to the limited workforce of Nectec, and given the impact of breakthrough technologies, the scope must expand beyond particular companies. Solutions need to come in a platform, offered as a basic service that the masses can use for free. Investors are needed to put money into the platform to serve individuals users and businesses.
"We very much hope that the Digital Economy and Society Ministry will be interested in these off-the-shelf technologies, as they require investment in infrastructure and cloud services," Mr Chai says.
To make an AI service platform requires investment by enterprises, and Nectec is talking to many companies, including Chinese firms, Thailand telecoms and other large corporations.
"To better serve the many sectors, Nectec has to change our role," Mr Chai says. "We would like Thai society to view Nectec as a countrywide technology platform. That means technologies that we already invested in must be a platform serving a great deal of users. AI-related projects will be lifted onto the platform, and there should be both free and premium services, making the system sustainable."
Having a broad range of users is critical to creating an ecosystem. In chatbots, for example, Nectec has talked to mobile operators so that the technology can serve a wide range of uses.
Nectec's goal is not to make money, but rather to create ecosystems for new technologies that people will want to use, platforms that can exist by themselves and become utilities of the country.
"That does not only cover software, but also hardware," Mr Chai says. For example, the precision farming system that Nectec has been developing has a chance for widespread use and a possible future as a Thai brand. It's a significant step, because the platform is made by Thais, serviced by Thais and used by Thais.
"I would like every high-precision technology to be like that," Mr Chai says.