Op-ed: What customers must know about “Know-Your-Customer”

Author: Sutapa Amornvivat, Ph.D.
Published in Bangkok Post newspaper/ In Ponderland column 20 March 2019

At the end of last year, Thailand’s cabinet has just approved a new legal draft in principle to set up the National Digital ID company (NDID). The company will build and oversee a digital platform that allows businesses and government departments to verify citizen identities against each other. This project sets out to be a key enabler for Thailand to move towards the 4.0 era.

This process that NDID will tackle is the process of identifying customers is called “Know Your Customer,” or KYC. Oftentimes, it is associated with banking and anti-money laundering regulations. However, it is highly relevant for many businesses, ranging from e-commerce, gaming, to social media with an underlying goal of protecting customers against identity theft and misuse of services. Despite being a less discussed topic, KYC is now at the very heart of the digital transformation which is all about designing products that put customer’s needs first.

NDID, as a product of well-coordinated effort between businesses and government, will enable us to transit faster to the full digital experience. Customers will be able to open a bank account online without visiting any physical branch, provided that they have verified themselves at another NDID member before. This can significantly improve ease of doing business in Thailand, boosting the much needed productivity for the economy. It provides a core infrastructure for the country, with a vision similar to that of Estonia’s X-Road, the basis of e-government, at least in the aspect of exchanging citizen’s identity information.

Why should we care as customers?

In an offline world, KYC is done by having customers meet company’s staff face-to-face so that the customer can be verified of his/her identity. As services move online, physical meeting is no longer a viable option. This friction can make or break digital transformation.

In Thailand, companies are improvising KYC solutions on their own without a standard solution. An online forum, Pantip.com, requires users to take a picture of themselves holding a physical ID card to create accountability for their posts and comments. Recently, Rabbit card, an e-wallet for Bangkok’s rail system, required users to register themselves face to face at train ticket booths; while a new digital bank, TMRW by UOB, asks customers to submit selfies and scan fingerprints at kiosks in their physical branches.

More private information is being asked to share through digital channels in the name of KYC. This translates to increasing risks of privacy breach and a loophole for abuse of personal data.

Customers should be aware of what we are giving up in exchange for services, and must demand to know what our private data will be used for and who it will be shared with. This should also be done through agreements or terms of services that are written not in legalese, but a readable language.

Data privacy and security of citizen’s information should top the priority in any implementation of an identity platform. NDID plans to use blockchain, a distributed ledger technology, touted as a highly secured way to store information given its immutable design. However, blockchain is not hack-proof. Several incidents of blockchains being hacked have been reported recently. Therefore, additional effort to ensure security and privacy is necessary to protect consumers and inspire trust among the public.

Another delicate point to address is on setting incentives for different participating parties. As the platform lets one party verify an end user against another party as a trusted source, there is an incentive for those with more identity information to withhold data, by design. This could be a roadblock in scaling the platform especially at the beginning. Re-aligning incentives to get those with data to share will be a difficult task. But its achievement can help build a strong foundation for small start-up companies in Thailand to gain access to this new system that solves the KYC issue.

The NDID initiative has a great potential to be the total solution for KYC standards in Thailand. The new infrastructure will provide a building block for open-data policy. Next step is to join force between NDID, government agencies as well as corporates. With NDID as a way to verify users and grant access permission, public and private parties can now open up data more using open API technology (as I wrote in this column in September 2018). Such technology will allow innovators to collaborate in building innovative services for the public.

This will put data back in hands of the rightful owners, the consumers, and empower them. The open-data policy will promote a culture of transparency and contribute enormously to the technology development in the age of AI.

Op-ed: Forget me (not)

Author: Sutapa Amornvivat, Ph.D.
Published in Bangkok Post newspaper/ In Ponderland column 23 January 2019

Why does one want to be forgotten?

Have you ever wondered if that embarrassing photo of yours from college days still exists in the vast World Wide Web? What if it was a comment on an online discussion forum falsely accusing you of a crime?  Since the rise of social media, we have been warned that everything you do online will likely be there forever, exposed to the public.

On top of your data exposed on the internet, recent scandals in 2018 revealed just how much companies are collecting your private data with or without your knowledge. The Cambridge Analytica incident shows how customer data can be shared and abused by companies.

With the arrival of the “right to be forgotten” in the EU’s General Data Protection Regulations (GDPR), you can now ask companies to delete such personal data.

So, what exactly is the right to be forgotten?

It is the concept that people should have the right to ask companies to delete data collected about them. The fundamental idea is that personal data belongs to people, and they should be able to delete it, as and when they wish.

Despite being a new buzzword, the right to be forgotten is nothing new. It has been officially recognized in France’s laws since 2010. In 2014, a Spanish man asked Google to delete information about him in the past, which appeared on Google search results; the Court of Justice of the European Union (CJEU) ruled in favor of the right to be forgotten based on what is implicitly suggested in the EU’s Charter of Fundamental Rights for European citizens. The ruling outcome mandates that Google must accept requests to delist websites to protect user privacy.

The renewed interest in the right to be forgotten came in 2018 as the EU’s General Data Protection Regulations (GDPR) took effect last year which made this right explicit. It is also one of the most controversial and difficult rules to comply.

Why is it controversial?

The Google Spain case sparked a major debate around whether this law is just, and rightly so. This debate is highly relevant for Thailand now as we will soon embrace the new data protection laws. Several key questions stemmed from the debate remain unanswered.

Firstly, is the “right to be forgotten” of an individual interfering with the “right to know” of the public? The flip side of erasing a piece of information is that it is no longer available to the public. Giving courts and governments the authority to delete data by exercising the right to be forgotten could lead to a slippery slope of power abuse (such as mass censorship). In the classic George Orwell’s nineteen-eighty-four, the fictional Ministry of Truth has the power to rewrite a history. Are we ready for the possibility of such authority? Moreover, making certain information accessible for some, but not others could worsen inequality among the public.

Currently, the data to be erased must be deemed “irrelevant, outdated, excessive, or inaccurate.” But who should be responsible for this decision? This blurred line can be a threat to freedom of speech. The rule of thumb for now is whether knowing a given piece of information is beneficial to the public.

Why is it difficult to comply?

Secondly, is it even possible to erase history? Suppose that it is deemed necessary to delete someone’s data. Stories are rarely about one person. It is not easy to delete information about one individual without affecting others.

Say, if a customer requests a bank to erase his data, deleting his profile would be the obvious step. The hard part is how to best handle transaction data. Should money transfers information to this customer be deleted as well? Should it be kept as is, kept partially (changed to money transferred to an anonymous person), or deleted entirely?

Imagine a web of information that links a large network of individuals. How do we deal with the situation when an individual’s data is embedded in a complex algorithm? How far should a company go to erase all traces of user history? Clear boundaries will need to be defined, if we were to institute such right.

The problem can get even more complicated in case of blockchain, the new decentralized ledger technology set to disrupt banking and many other industries. The core promise of blockchain is that once a block of data is added to the chain (the ledger), it cannot be altered nor deleted. This immutability makes the system trustworthy. New research is being done to address the right to be forgotten in blockchain, but this could defeat blockchain’s purpose. As many companies are adopting the technology, the tension with the new laws can only multiply.

Are companies prepared?

Heavy fines and penalties are at stake. Failure to comply with the GDPR, including the right to be forgotten, can result in a fine of up to 20 million euros or 4% of worldwide annual revenue. Some countries such as Italy, Austria, Germany, impose an additional penalty of imprisonment for company directors.

Based on the current trend, it is possible that the right to be forgotten will soon arrive in Thailand. Companies that are collecting and processing customer data should be prepared. The first step would be to recognize the importance of customer privacy and understand exactly what kind of customer data you are collecting and processing.

The concern for data privacy will become increasingly significant especially when companies try to strike a balance between data privacy and optimal user experience design. Too strict of a law could be a double-edged sword. One clear example is that now we are forced to accept these long cryptic terms of service, instead of seeing what is really important.

Ensuring the right to be forgotten represents a step forward for privacy protection in this age of technology. However, in practice there is still a considerable gap between concept and implementation. Moving forward without addressing the complex reality of implementation would mean the risk of taking one step forward, but two steps back.


Op-ed: Who’s Afraid of Peer-to-Peer Lending?

Author: Sutapa Amornvivat, Ph.D.
Published in Bangkok Post newspaper/ In Ponderland column 29 November 2018

The recent announcement by the Bank of Thailand on peer-to-peer (P2P) lending rules represents a significant paradigm shift. If all goes to plan, by early next year Thailand will be among the few Asian countries, most notably China and Indonesia, to legalise this match-making platform between lenders and borrowers.

While this move speaks volumes about the progressive policy environment for Thai fintechs, the risk manager in me says that we should carefully weigh the pros and cons before jumping in.

The concept of P2P lending is a platform as a matchmaker role between investors — retail or institutional — with available funds, and borrowers who need the funds. In theory, such a platform can greatly enhance market efficiency and better allocate funds between people, overriding cumbersome intermediaries such as banks. It would be a win-win situation for both sides — investors get better returns; while borrowers get better interest rates and access.

Yet, this year, we have all heard of a series of bad press in China about hundreds of P2P lending platforms having gone bust. Its repercussion of eroding trust in this industry has become the key debate in the global fintech space.

On the other hand, success stories rarely make the news. There are P2P lenders in the US and the UK that have performed well for over a decade. Even in China, some have emerged as victors despite the growing mistrust. Thailand is blessed with the opportunity to learn from these experiences to ensure we start off on the right footing.

There are key lessons to learn from the crisis in China for all parties — those thinking of starting up a new P2P platform, potential investors, consumer protection groups, as well as regulators.

First and foremost is how to mitigate fraud and security risks.

In China, before the recent crackdown, there were several cases of outright fraud by platform operators. Ezubao, one of the largest players, ran a Ponzi scheme that made off with 100 billion yuan (nearly half a trillion baht) from investors. Many others were caught in a similar fashion. Regulators should have a stringent monitoring system to prevent such scams. Investors, too, should be wary of the trustworthiness of P2P platforms.

In terms of security, given the popularity of the P2P lending market in recent years, this industry will surely continue to be a top target for cybercriminals using stolen identities to create loan applications with synthetic credentials. Platforms that place an emphasis on cybersecurity and airtight KYC technology will thrive without risking fraud losses or exposing investors’ and borrowers’ sensitive information. Protecting the platform against fraudulent activities, of course, is not cheap. Therefore, finding the right trade-off between platform fees and fraud risks will be crucial to borrowers and investors alike.

Another key lesson is how well P2P platforms manage liquidity risk. Most of the recently collapsed platforms in China like PPMiao (which prompted a major protest in Beijing) experienced a run, similar to a bank run, where investors withdrew money at the same time, leading to a liquidity shock. This can be mitigated by having a separate fund to buffer such a risk or a fund portfolio mix of some larger investors similar to platforms in the US and UK.

Last but not least, we must educate the public. Borrowing money from friends and family is a simple concept, but with a digital platform and internet connection, P2P lending broadens the reach for borrowers to access capital from someone else.

At the same time, investors have options to seek higher returns than bank deposits. In this process, the platform has a role to play as a matching mechanism. It should possess a robust and time-tested credit scoring model allowing investors to understand the risk and make better informed decisions.

Underwriting a loan, especially one without collateral, is not easy and requires deep consumer credit knowledge. Positive investor returns are never guaranteed. Investors in P2P platforms must understand the risks of this new kind of investment. A misconception exists that P2P investment is very low risk. This has led investors to discount the importance of diversification. As seen in China, retail investors lost their life savings by mistaking P2P investments as safe deposits with promised returns; altogether too good to be true.

P2P lending has the potential to address the loan shark problem in Thailand and improve financial inclusiveness. It could also raise awareness of the Thai public about financial investment to think about returns on their financial assets beyond bank deposits, preparing us to enter an ageing society.

No one should be afraid of P2P lending. If done right, the introduction of P2P will no doubt present many opportunities for Thailand. However, we should tread carefully. It is vital we learn from the past lessons — both the successes and failures of other P2P platforms to ensure their best practices are emulated and their mistakes are not repeated.

Op-ed: Important lessons from a chatbot developer

Author: Sutapa Amornvivat, Ph.D.
Published in Bangkok Post newspaper/ In Ponderland column 31 October 2018

The smoother and deeper our conversation with Siri, (or with Alexa, Bixby or Cortana for that matter) the more realistic Dan Brown’s Winston in Origin becomes. Many experts are predicting that demand for new mobile apps is slowing down with the next trend for user interface being chatbots. Of course, much more work in conversational AI will be necessary before we reach Winston’s level of chatbot advancement. At least now, users are becoming more familiar and as a result, these interactions provide a constant source of data for machines to “learn”.

Many retail companies are utilising a chatbot platform for various use cases ranging from customer support to recommendations of products and services. According to Oracle, 80% of businesses polled in France, the Netherlands, South Africa, and the UK last year said that they will resort to chatbots by 2020. In fact, one French-based multinational beauty chain already has three chatbots, one to assist in makeover reservation, one to help customers match colour shades, and the other to offer makeup tips and how-to videos.

It is no surprise that Thai businesses are jumping on the bandwagon, especially with the Thai population being so reliant on chat apps that we frequently chart in the top 10 for global social media usage. When such clear business value is involved, the rate of developments in Natural Language Processing (NLP) techniques within AI can be expected to accelerate. Many cloud services are already expanding to cover Thai language and offering standard tools to guide non-programmers to build chatbots themselves.

We at SCB Abacus have launched a number of chatbots this year. One of our first chatbots, “Puek Hom”, is an expense tracker chatbot on Facebook Messenger. We learned that Puek Hom is more versatile than expected, attracting not just millennials but also Gen X users. This gives us hope in the potential of chatbots within the competitive market of personal finance. The most recent chatbot, “Perm Poon”, developed for SCB Asset Management, can now handle lengthy interactions and advise on complex products, such as mutual funds.

Our experiences in developing these chatbots tell us that making a great chatbot is more of an art than science. Here are five things I want to share with Thai corporations looking to build chatbots in the future.

A great chatbot needs a persona.

You need to do research and understand target persona for an effective communication style. Diverse skills are required, such as user experience and content design, not just software development. All these skills are an essential part to a user-centric product design because they ensure that you serve your users’ best interests. Humanisation is even more crucial — chatbots should be able to replicate how your best agents interact with different types of customers, especially as the bots may not always know what to answer.

A great chatbot is personalised.

A great chatbot learns from each individual’s conversation history and other peripheral data and responds accordingly in real-time. Imagine a food delivery chatbot that can recommend restaurants based on previous chat history, previous orders and current location, instead of asking users to go through the same back-and-forth chat every time. Building a chatbot capable of personalisation requires technical knowledge of database architecture, and the ability to integrate internal and external data systems.

A great chatbot must be able to hold a conversation.

To keep users coming back to use the chatbot, it is crucial to keep the conversation flowing and engaging. This means that, like in a good human-to-human conversation, there is never “dead air”. One way to do so is by adding value to the conversation by anticipating users’ needs based on previous chat history, without being explicitly asked. A seamless experience can also help. A good chatbot should have omni-capability where conversation flows seamlessly across channels.

A great chatbot should know its own limits.

Having a chatbot for customer service is quite well-received globally, with 45% of users around the world preferring chatbots to humans, according to a 2017 poll by a market research firm Grand View Research. However, this will differ from industry to industry, as it can be difficult to create a personalised experience for certain products. So where do you draw the line? It goes back to the first point I made about knowing your target audience. Often it can be difficult to know for certain before the chatbot launches. Therefore, one must have be quick to reiterate the chatbot in short periods to retain and engage customers.

Developing and maintaining a great chatbot is a long journey.

Like other AI products and especially NLP tools in Thai, a chatbot needs continuous improvement and huge data training process to enhance its ability to handle various types of conversations. It is also important to note that a chatbot cannot solve everything. Companies such as Amtrak has created an additional chat-tree where, if reached, users can get in touch with a real customer service representative for questions that require human attention.

A great chatbot is a significant investment in terms of cost and time. It is arguably more beneficial to companies with large customer bases, but it can also benefit smaller companies looking to expand from niche to mass markets.

Not only will we see chatbots becoming the interface of choice, but they will also continue to disrupt our search behaviour. Currently, Siri and its Android counterparts are capable of offering a list of general information, but lack domain expertise to provide “the one right answer”. On the other hand, chatbots that businesses build will possess such domain expertise, but how do we ensure that those chatbots are discovered?

Instead of having to browse through a list of information provided by Siri, perhaps we will see a future where multiple chatbots will collaborate with one another to deliver that one right answer straightaway.

Op-ed: Moving towards digital government

Author: Sutapa Amornvivat, Ph.D.
Published in Bangkok Post newspaper/ In Ponderland column 20 September 2018

I recently joined a panel discussion at the Ministry of Commerce’s forum on technology development and connectivity in the CLMVT (Cambodia, Laos, Myanmar, Vietnam, Thailand) region. Talking about this topic often brings us to physical connectivity like road and rail; while we overlook another important aspect of connectivity especially in this era — the flow of information or data across countries. This type of connectivity forms a crucial part in supporting sustainable development and shared prosperity.

The key enabler for such connectivity requires government investment in its own digital infrastructure that will provide a strong foundation for people and business to build upon.

One particular area I want to discuss is digital infrastructure for data sharing. Given the massive amount of data at hand, governments can provide a strong catalyst to advancement in technology like AI and big data by taking on a role of data aggregator and a platform that gives data back to the rightful owners, its citizens.

A leading international example is the case of Estonia, a small country in Europe that used to be a part of the dissolved Soviet Union. The country has transformed itself into a world leader in digital government, especially for its data infrastructure.

The Estonian government has invested in a secured data exchange platform called “X-Road” that allows residents, public institutions and private companies to share information. Participants will gain access to e-ID, a system that verifies digital identity. In essence, this is a public data sharing project that connects civilians to the government database containing secured digital IDs. It allows Estonians to access digital services securely, ranging from tax filing and reviewing medical records to selling cars. This paperless model is estimated to save 2% of Estonia’s GDP each year, which amounts to US$500 million.

Other countries have followed this model. In early 2016, Japan launched a digital identification programme called “MyNumber”, which helps residents access and manage a range of government services online such as social security and income taxes. Finland, Oman, Azerbaijan and Palestine have already adopted a similar system. Others have shown interest with Germany’s Chancellor Angela Merkel visiting Estonia’s “e-government showroom” in 2016.

In building such a system, data privacy is at the heart of this matter. Proper data governance must be put in place to ensure transparency. By logging into Estonia’s State Portal, residents can check who holds their information, who can access it and who has accessed it. A separate entity, the Data Protection Inspectorate, takes the role of regulation enforcer. The system is designed to allow consumers to have firm control over their own data.

In Thailand, progress in terms of government digital infrastructure is under way. At least, by October this year, all government agencies will no longer ask for a hard copy of a national ID card. A copy has long been required to do any official business, from getting a driver’s licence to registering a company. This is a great step in the right direction.

As a part of the national strategy, the prime minister has announced a national agenda, the Digital Government Development Plan, to be spearheaded by the Digital Government Agency (DGA, previously Electronic Government Agency).

One of its earlier initiatives is to build the Government Cloud (G-Cloud) that will provide cloud computing and storage services for government agencies. Several offices such as the Electronic Transactions Development Agency and the Ministry of Public Health, have already started using its services. This could be a foundation for a backbone data exchange system for Thailand, like X-Road in Estonia.

For the next step, the government should open up its cloud system to non-government parties. Open APIs would allow any developer to connect to the platform and access data (with the consent of users) so that they can build innovative services for the public. Imagine a “ThaiGov for Developers”, similar to those offered by Facebook and Google. An open system would encourage the participation of creators and foster a collaborative culture.

With such a data exchange centre, we will be giving back the right to control personal data to the people. A greater sense of ownership should lead to more vigilance over their data, regardless of whether it is collected by the government or private parties.

Undergoing a digital transformation is not easy. I have witnessed first-hand the banking industry’s continuing struggle through this process. The strategy is more than just a matter of technology. It is also about culture and ways of work.

Open data by the government and a public platform that gives data back to the people will be a step towards creating a ripple effect in terms of data culture. It could also be a big leap forward in educating the public about data privacy.

Op-ed: Thai Healthcare in the Digital Age

Author: Sutapa Amornvivat, Ph.D.
Published in Bangkok Post newspaper/ In Ponderland column 15 August 2018

A few weeks ago, I participated in a workshop to rethink the master plan for the Thailand Centre of Excellence in Life Sciences (TCELS), the committee of which I serve as a director. This reaffirmed my belief in Thailand’s potential within the medical sciences and healthcare industry. As a country, we are already renowned as a medical service hub of Asia. With recent advancements in technology, we can further build on the strong foundations especially in the area of research and development (R&D).

The major breakthrough that could leapfrog the healthcare industry is the combination of two key technologies: Internet of Things (IoTs) and data analytics. These technologies are significantly complemented by the aggregation of data which provides a great source of information for medical research. These advances will have a resounding impact on modern society, and deserve consideration by the highest level of decision-makers.

The first key enabler that has fuelled the boom in data technology is the capability to collect new data, anytime, anywhere at a reasonable cost. More data on nearly every aspect of our everyday lives are constantly being collected than ever before.

This is made possible from advancements in sensor technology.

Today, it is likely that you are already walking around with a bundle of sensors. There are approximately 24 million smartphones in Thailand; each of which is equipped with an accelerometer, gyroscope, magnetometer, and several other sensors embedded in a tiny chip to collect our data and facilitate interactions with the phone.

The most widely adopted IoTs today, aside from smartphones, are perhaps wearable fitness trackers such as Fitbit and Garmin that can typically monitor movement, sleep cycles, and heart rates. The market for IoTs, especially in healthcare alone, is expected to reach US$163 billion within a few years from now, according to research by Accenture.

There are also advanced sensors dedicated for specific purposes, especially for monitoring of patients with chronic diseases such as cancer and diabetes. For example, iTBra is a dual breast patch that can monitor circadian metabolic changes in heat, correlated to cellular activity common in breast tumors.

Such devices allow for monitoring of our health data like never before. Moreover, as the size of these wearable devices shrinks even further to something implantable, injectable, or swallowable, they will become a part of daily life more easily. Earlier this year, scientists from China’s National Centre for Nanoscience and Technology and Arizona State University jointly developed nanorobots that can enter the bloodstream to target cancerous cells.

In addition to the capability to collect data, advancements in data analytics are another key enabler that will help revolutionise the healthcare industry.

Artificial intelligence (AI) algorithms can find meaningful patterns from data streaming in from these devices. By detecting anomalies, AI can predict and warn users to take actions pre-emptively. These changes in technology trends will allow us to move from a “sick-care”, where we respond passively to our symptoms, to true healthcare.

Aggregating data will be crucial in connecting the dots for better healthcare. At the individual level, Apple’s Health app on iPhone demonstrates an example of data aggregation from its built-in sensors and IoT devices and presents the data in a way that help users engage in their own health and well-being. Healthcare providers, including Thai hospitals, are moving towards better data aggregation with Electronic Medical Records (EMR).

Beyond the individual level, aggregation of health data can lead to strong value creation especially in research. For example, accumulated data from patients can be used to reveal information for drug discovery. Powered by AI algorithms, a pharmaceutical start-up, Berg, analyses the differences in patterns between mutated cells and healthy cells among cancer patients to develop a cure for pancreatic, breast and liver cancers.

Such tasks are not possible without massive amounts of data from many patients. Imagine what a well-structured data-sharing system for health research can offer. The Thai healthcare/medical research community will benefit greatly if these initiatives are implemented.

The sharing of health data will be key to future breakthroughs that can transform healthcare industry. The data, once anonymised and secured properly, should be thought of as public goods. We can focus on the technology that allows easy anonymisation and security such as blockchain.

Governments can also play a role of data aggregator by providing infrastructure. Initiatives like the government cloud can promote collaboration between healthcare providers, patients, and researchers into fruition. Certainly, the use of new technology does not come without risk and challenges. Laws regarding data privacy and security must be in place.

Last month, Singapore has just suffered a healthcare data breach of 1.5 million records. SingHealth, a major governmental healthcare group, was infected with malware from a single office workstation, which allowed the hackers to access the entire healthcare database. This is a wake-up call to companies who own sensitive customer data to step up measures in cyber-security.

After the incident, SingHealth imposed additional security measures such as Internet Surfing Separation — where internal staff networks are isolated from the wider internet which reduces the exposure of government healthcare data to cyberattacks. This case serves as an important lesson but should not be considered as a roadblock to long-term technological progress.

With more shared data, researchers can ask better questions and discover better solutions to our health problems, leading to an era of proactive healthcare.

Improved healthcare data infrastructure will benefit all players in the medical/health ecosystem — not just patients, but also doctors and researchers. By allowing more access to these data given proper governance, we have a chance to attract the top domestic and international talent to the industry. We need to admit we should not be too Thai-centric, and admit a talent shortage. A strong technological foundation will prove crucial in enticing the best professionals, and further expanding Thailand’s position beyond the medical service hub of Asia into a research hub as well.

Op-ed: Putting the ‘AI’ in Thai agriculture

Author: Sutapa Amornvivat, Ph.D.
Published in Bangkok Post newspaper/ In Ponderland column 20 July 2018

Last month at the SingularityU Summit, I was exposed to some of the latest and most exciting technologies across various sectors. Of particular interest to me were the technological developments within the agricultural sector — an industry so vital for Thailand.

How can artificial intelligence (AI) be leveraged in agriculture? And more importantly, how can Thailand benefit from these advancements?

Agriculture has always been one of the core economic activities in Thailand, with over 40% of Thai workers employed in the industry according to the National Statistics Office. Yet, this sector contributes only 10% of GDP, and is in decline. Furthermore, mechanisation for harvesting, processing or irrigation requires investment and has been sparsely adopted by Thai farmers.

But with the changing technological landscape today, perhaps we can be more optimistic. The cost of technology has dropped drastically while productivity gains from adoption have risen. The agricultural sector could leverage new trends like AI, that surpass the previous linear evolution of mechanisation. Combined with the abundance of local wisdom, AI can play a part in producing more precise farming that leads to exponential agricultural growth.

To make this a commercial success, a holistic approach must be considered. Firstly, new technology should be applied to improve efficiency in agricultural production. This is the most obvious use of AI for agriculture — to make farmers more productive, and the business less labour-intensive. Some of the AI and Internet of Things (IoT) technologies are already being successfully utilised in developed nations. For instance, an agriculture drone presents an effective method for collecting data from an aerial perspective. When equipped with software programs, it can analyse data for potential improvement in crop yields and farming efficiency. These AI tools can also be applied to livestock. In the case of “Cow Fitbits”, AI can learn behavioural patterns in order for farmers to better cater to animals’ wellbeing.

Beyond efficiency gains, AI systems can help safeguard agricultural activities against various risks. Many technologies now offer precision farming solutions that control unpredictable situations such as pests and diseases. One start-up called Blue River Technology helps farmers detect weeds by deploying drone-based phenotyping technology to identify which plants do not belong in the field. In addition, farmers can also choose to mitigate risks by purchasing agricultural insurance that is underwritten using AI and big data, to allow for a better cost structure and secure the necessary funds should unexpected events occur. Lastly, with the prevalence of internet and social media usage, communities around the world are becoming more connected. Shared digital platforms allow farmers from around the globe to exchange knowledge and best practices. Aggregator platforms like Aggrigator, where farmers are constantly synced with market clearing prices, can help to identify niche demand in real-time. Social media also allows for more fruitful rural-urban interaction which will, in turn, foster new innovative business models.

How do we go about implementing this in Thailand, exactly?

A key element to successful adoption is to foster an innovation ecosystem within the industry. The trend of urbanisation has resulted in much of the younger demographic migrating to the cities in hopes of better prospects. Successful modernisation through technology will require the right talent. Unfortunately for many, agriculture is not perceived as lucrative or attractive.

But now the tide is shifting. We are starting to see a new generation with an entrepreneurial mindset returning to their families’ farms or starting anew. Stronger networks will also help: Over the past four years, more than 7,000 young individuals have joined the Young Smart Farmer programme under the supervision of the Ministry of Agriculture and Agricultural Cooperatives. The programme is expected to attract an additional 6,000 this year alone. More initiatives like these are emerging and, when coupled with network effects, can lead to more talent re-entering this industry.

This new generation has already brought novelties to their farming practices and channel of sales, experimenting with irrigation technology and Facebook advertising. Thai startups have also entered the agriculture space to help farmers connect with technology. For example, Ricult, co-founded by a young Thai entrepreneur, has developed a satellite imaging solution to assist farmers in planning crop rotations. These innovators would certainly be more open to embracing AI and champion this sector.

Successful implementation of agricultural transformation will depend heavily on the support and commitment of the public sector. A state-level agricultural transformation will be extremely costly with many trade-offs; therefore it is important that decisions should not be dictated by politics but guided by evidence and data.

Indeed, the government has an active role to play in creating a data-driven environment. A centralised agricultural data platform will allow AI tools to be deployed. This data platform should systematically collect and continue to update farming data such as crop seasonality, pest issues, soil conditions, weather forecasts as well as local knowledge across areas and crop types. The government should take the initiative not only in creating the centralised platform, but also in playing a field agent role to collect agricultural data and accumulate industry knowledge. This will allow technology providers to access readily available data and offer localised services to farmers more cheaply.

Without a doubt, this will be a huge investment from the government, requiring strong leadership and vision from the top. Indeed, applying AI in agriculture is a long-term game that will involve sub-par growth initially and many uncertainties. On the bright side, the advanced technologies that will uplift our agriculture sector already exist and are in use around the world. The real challenge now is to align our priorities and maximise our readiness to successfully adopt these technologies.

Financial services, telecoms and manufacturing are currently leading in the AI race. Thai society cannot afford to leave behind one of our most crucial and socially impacted industries: agriculture.

Data Science Workshop at Mahidol Wittayanusorn School

On the 7th July 2018, SCB Abacus hosted a Data Science Workshop for the students at Mahidol Wittayanusorn School. The Abacus team led the workshop in teaching the students on process of creating of an interactive chatbot.

Students were tasked with designing the purpose and objective of their own chatbot in order to solve a problem that they experience or observe around them. Beyond just the technical aspects of the workshop, a significant emphasis was placed on the design process to encourage creativity and critical thinking.

It was a very enjoyable day for both the students and the team. This workshop is just one of the many initiatives by SCB Abacus to inspire the next generation of technologists as well as furthering the knowledge base of the AI community in Thailand.

Stay tuned for more events like these in the future!


The contents of the lesson can be found at this Google docs link:




Op-ed: Educated by Artificial Intelligence

Noted speakers at the Singularity University Summit last week included Sutapa Amornvivat (second from left), author of this article and a regular Bangkok Post contributor. (Photo via SCB Abacus)

Author: Sutapa Amornvivat, Ph.D.
Published in Bangkok Post newspaper/ In Ponderland column 27 June 2018

Last week, at the SingularityU Summit in Bangkok — a two-day event that comprised a series of talks by visionary technologists inspired an audience who were mostly business leaders and key policymakers, I was honoured to join Dr Vivienne Ming of Socos Labs and Dr John Jiang of CP on stage to discuss Artificial Intelligence (AI) and its impact on Southeast Asia.

The recurring theme throughout the event was the concept of “exponential” technology. This term emphasises the rapid growth of impact from breakthrough technologies. By nature, humans are linear thinkers. We often (mistakenly) over-estimate the impact of technology in the short run, but underestimate its effects in the long run. As a result, we can dismiss new technology prematurely, and thus forgo what could become the next big thing.

With such an exponential trajectory, especially in AI development, the future of work is changing rapidly. Many fear that AI is a very real threat to our job security. As a parent of two preschoolers, I have always pondered what skills our children will need in the next decade.

It would have been similarly difficult for our parents to guess that AI technology would become such an integral part of our present everyday life. My fellow speaker, Dr Ming, advocated that we should not be guessing at all. Instead, she suggests that the best way to prepare our children is to train them to embrace the uncertainty of the future — to be creative, adaptive and problem explorers. This requires a different kind of learning experience than the current education system offers.

Can AI, despite posing a risk of job losses, also enrich education for the future?

The answer is yes, as we have seen tremendous progress in AI that can play an important role in solving pain points of students and educators.

Market research by TechNavio forecasts that the use of AI in the US education sector will grow 48% annually from now until 2021. Tech giants like Facebook, Microsoft, IBM are also paying attention. Last year, Facebook, in support of the Gates Foundation, announced a partnership with a charter school network, Summit Public Schools, to develop a learning platform powered by AI algorithms.

Currently, the most prominent use case of AI in education technology is to create a personalised learning experience.

The idea behind personalised learning is that individuals develop at different rates, with varying proficiencies and interests. In Thailand, the effort to “personalise” has led some schools to sort students into classrooms based on performance. There is a downside to this policy. Those with poor performance could continue to spiral downwards when they are grouped together. Research shows inconclusive evidence whether this initiative positively impacts a student’s academic achievement.

This issue is particularly important in Thailand where educational inequality is high. According to the Thai Office of the Basic Education Commission (Obec), the average size of a public secondary school classroom is as large as 50 students. For OECD member countries, the average class size is 24 — less than half of Thailand’s.

This is where AI can come in.

Based on a student’s own pace and interests, AI systems can custom-design individual learning experiences, so that each student can find the most optimal path towards their own learning goals.

Imagine a classroom where each student can move through class materials at their own pace. UCL Knowledge Lab in London built a software called AIAssess. The system assesses a student’s understanding of a concept, and dynamically adjusts practice problems to match that level.

During the class, Lenovo’s virtual classroom uses computer vision to help teachers interpret students’ gestures to identify those lagging behind and in need of special attention. Facebook built an app that helps student create course plans to best achieve their learning goals.

After class, a new teaching assistant, Jill Watson, who shares the same name as IBM’s AI platform, helps teachers respond to student questions. A team of scientists at the Georgia Institute of Technology built this machine using data from online discussion forums. Using advanced Natural Language Processing algorithms, it can identify common questions and mistakes that students make.

At home, students can read a personalised textbook created by Deep Learning algorithms. A startup company, Content Technologies, uses algorithms to condense books into short summaries and identify key materials. In digital format, we can track individual student interaction with books such as time spent reading each paragraph, or whether they are stuck on difficult topics. Given such data, AI can help identify students’ pain points and appropriately respond to help improve learning.

The use of AI in education is still at its early stage. Personalisation by AI combined with online open course platforms like Khan Academy, Coursera, and EDX will create a truly customised experience. We might see an entire degree composed from online materials from different sources.

By adopting AI technology to personalise the education system, we will be able to shift the focus from students aiming for standard degrees towards a more wholesome learning experience. This will make our children better equipped for the unknown future.

While AI does indeed offer limitless potential for human progress, we should be conscious of its impact on our mentality. One of the many things that keeps me up at night is how AI technology can weaken the human mind. For example, we can no longer remember many phone numbers, directions, or even tomorrow’s calendar. The argument goes we are then free to engage our minds in other creative pursuits or solving more complex problems. But do we really?

The last chapter of Jerry Kaplan’s book, Humans Need Not Apply, suggests that the future might not be such a dramatic humans-vs-machines scenario. Humans are already ceding control to machines willingly. Nowadays, many decisions are automated; drivers trust Google map for directions, and doctors depend on machines to test, diagnose and treat patients. That’s the benign-looking threat Kaplan speculates will happen. We will welcome computer colonisation.

Let’s not allow our laziness to be our downfall!

Seismic shift reshaping data landscape

Bangkok has more Facebook users than any city in the world, and Thailand must begin privacy laws. (Main graphic clipped at Slideshare.net)

Author: Sutapa Amornvivat, Ph.D.
Published in Bangkok Post newspaper / In Ponderland column 16 May 2018

Over a month has passed since the Facebook scandal which has sent shockwaves across the tech industry and around the globe. Facebook came under fire following an incident involving its user data exploited by Cambridge Analytica, a UK-based analytics company that was contracted in US President Donald Trump’s election campaign.

Facebook, as a platform that allows data harvesting to happen, faced a strong backlash from the public and lawmakers. In its defence, the company does have terms of service that prohibited the actions undertaken by Cambridge Analytica.

Following the scandal, Mark Zuckerberg, the founder and CEO of Facebook, testified before US Congress. From a public relations standpoint, the Senate’s lack of computer literacy seemed to give Mr Zuckerberg the upper hand. Many of the members did not understand the basic business model or functions of Facebook.

But regardless of the public outcome of the hearing, the fact of the matter is: Facebook did not go far enough in policing how companies were using data from its platform, an issue Mr Zuckerberg promised to address.

A month has passed since the hearing, so what has changed?

Facebook recently hosted the F8 developer’s conference, where they announced one of their most socially ambitious projects, a dating app.

This could be interpreted as a sign of confidence. They are willing to bet on dating — one of the most sensitive and private activities people engage in. Perhaps Facebook believes they still hold users’ trust over their data. In spite of the #DeleteFacebook movement, statistics have shown that very few people have actually deleted their accounts and left the network.

Cambridge Analytica has fared much worse, even filing for bankruptcy this month. It claimed the scandal has driven off its customers.

The public and legal spotlight should now return to these self-regulated platform-based companies. It has often been their prerogative to draw a line between what can be done with their users’ data. Given how events have transpired, data companies will have to demonstrate greater transparency and accountability in order to uphold the public’s trust.

To combat the issue of privacy, we should focus on protecting the public against abuse as in the case of Cambridge Analytica, rather than prohibiting data collection itself.

In the near future, we will see more data protection laws like the EU’s General Data Protection Regulation (GDPR). These are designed to protect consumers from organisations that process their data by ensuring greater controls and transparency. For example, the GDPR enforces the “right to be forgotten” whereby an individual can demand that their data is deleted after fulfilling the purpose when consent was given.

What are the implications for Thai businesses?

Once the GDPR comes into effect, Thai companies involved in the handling of data with EU partners will have to ensure compliance if they are to continue their business activities.

The issue of data privacy is particularly relevant for Thailand. Bangkok now has the highest number of active Facebook users in the world at 30 million. The country has over 46 million registered Facebook users, with the average Thai spending almost three hours a day on social media (the discrepancy in user numbers is partly attributed to a trend of people opening multiple accounts).

Either way, Facebook is undoubtedly an integral part of everyday life for many Thais. The three most popular platforms here — Facebook, Instagram and YouTube — all collect masses of data on their users.

Despite that, Thailand has no specific data protection law in place. Legal enforcement can only be taken when damage is inflicted on a victim as a result of a data breach. Moving forward, the government should strike the right balance between an internationally-accepted data protection regulation and one that does not impose any unnecessary burden on businesses.

Meanwhile, the government has drafted the Personal Data Protection Bill (PDPB), which has been revised over many years. The amended 2018 PDPB is still pending review by the cabinet this month, with no official timeline announced. It is clear that social media penetration has occurred quicker than the privacy laws can keep up. The Facebook scandal signals the need for such protection now more than ever.

But will legal measures alone be sufficient?

After all, it’s the users who consent to giving up their data. In fact, a university campus study in the US has shown that 98% of college students were willing to relinquish their private data in exchange for a free pizza. For many, giving up some personal information is a small price to pay for a free service or smooth user experience. Regulation alone may not be sufficient.

Promoting web literacy can provide another layer of protection as it helps people understand when their data is being collected and how it can be used. The public will become more aware of the implications when registering for Facebook quizzes such as “Which character are you in [the hugely popular Thai soap] Love Destiny?”. While seemingly harmless, some of these popular Thai Facebook apps have the same data harvesting capability as those used by Cambridge Analytica.

Web literacy should make people pause before taking that online quiz, or posting their national ID and bank account numbers as Facebook comments to online merchants — all practices that are common in Thailand.

The complicated issue of data privacy demands an equally comprehensive response from all parties involved. It will take time for the data industry — companies, regulators and consumers, come to terms with the right balance between privacy and better services.