The introduction of pioneering your enterprise with AI
We are here today with technology developers and organizational leaders to share our attempt in pioneering AI in Thailand.
“It is not just about Artificial Intelligence (AI), but also about transforming business strategies, value propositions, and business processes. AI is merely a strong facilitator; a very powerful tool.” Mr. Voradis Vinyaratn, Executive Director and Acting Managing Director of TCC Technology Co., Ltd. stated in an opening remark to launch TECH SPARTAN 2019: Pioneering Your Enterprise with AI. The event was organized by OPEN-TEC, inspired by T.C.C. Technology, held on February 22, 2019 at C asean. Mr. Voradis opened up his introductory speech raising three open-ended points of the day’s discussion around the topic of Artificial Intelligence.
The first open-ended point Mr. Voradis highlighted is the fact that today, pioneers are pioneering the adaptation of Artificial Intelligence or AI, leading the way to learn the early lessons of what worked and what did not work. Talking about what did work, there are many enterprises that are adopting AI technology on a wide-scale level across their industries. Especially for early adopter manufacturers, AI enables organizations to better orchestrate analytics, businesses, business intelligence, and real-time monitoring, and to be able to achieve fast growth.
On the other hand, it does not work to create technology without a strong underlying value-chain. We need to be clear on why we need the technology, and how to leverage business capabilities. Is it to achieve cost optimization, business process optimization, or to generate more revenue?
The second key open-ended point addresses the fact that ‘pioneers are deepening their commitment in AI.’ According to Mr. Voradis, AI pioneers are committed to giving life to their platforms through increased investments. With AI, businesses can be more predictive of their clients’ behaviors, with complex algorithms allowing them to be able to predict their clients’ decisions. From that point businesses will choose to harvest data in the right ways. According to Sloan Management and BCG Consulting report – based on survey over 3,000 managers and executives, the top 18% of AI adopters are devoting over 70% of their efforts to devise a new strategy for driving revenue and sales growth. In fact, over 91% of all enterprises expect that AI will be able to deliver new business growth by 2023.
Mr. Voradis highlighted the third open-ended message before giving the stage over to the day’s keynote speakers, the question of how we can succeed with AI technologies personally. The answer to these questions lies in creating connections that amount to a sustainable ecosystem. AI can allow us to light up hidden insightful data, extracted from the current immense amount of data and information. This intelligence is being developed rapidly and aggressively. The question is, how do we make use of these capabilities?
Genpact’s survey revealed that people between the ages of 18-37 are only fairly comfortable sharing their personal data with businesses, which indicates that we are facing challenges of rising concerns for data privacy, leading to the utmost concern over the ‘ethics of AI’. These are the key issues that will be discussed by the keynote speakers presented throughout the event.
Ethics of AI: Defining How We Should Use Artificial Intelligence
The first keynote speaker at TECH SPARTAN 2019 was Michael Araneta, Associate Vice President of IDC Financial Insights Asia Pacific. He addressed the issue introduced by Mr. Voradis concerning the ethics of AI.
Before going into the world’s rising concerns with the ethics of AI, Mr. Michael revealed in his presentation that today there are a lot of debates on how AI could potentially bring about disruption, and how AI might create unfairness or bias across societies. With these pressing issues rising, this is also a once in a lifetime opportunity, and it is important that we define how we should deal with AI in societies. Mr. Michael started his keynote with the fundamental theme of how using AI is related to the digital context of Thailand citizens within the increasingly growing digital transformation.
AI is everywhere, and in particular it is becoming more personal and more prominent in our daily lives. Mr. Michael gave an example of how Netflix and other digital platforms have the ability to know our behaviours, and suggest the next movies or shows, based on the information that they have from our previous interactions. This indicated that AI is being integrated into our lives on a more personal level in the forms of recommendation engines, chatbots, and professional tools, tied to how we enjoy various entertainment experiences throughout various platforms.
IDC is currently working on trying to define what AI really is and its implications, which was discussed throughout the event. The point that Mr. Michael emphasized is that AI is not necessarily just one single solution, but rather a conglomeration of solutions from different vendors, tools, and service providers that we are adopting. AI is a part of the continuum of artificial intelligence capabilities, but also automation capabilities.
Intelligence Automation Continuum: transformational tactic
Mr. Michael then went on to explain the progress of the intelligence automation continuum, starting from
- the Robotic Process Automation (RPA): at this point, the decision-making process is still defined by humans and may not yet be at the artificial intelligence level, with this a preparatory phase before artificial intelligence. IDC primarily looked at the data structure and the technology spending patterns of the institutions in Thailand, which revealed that there are a lot of ongoing RPA projects. We are able obtain data from a source and process documents faster by using implications in data patterns to support the process. Although there are no self-learning capabilities in RPA, the huge sets of data gathered and being processed is a good preparation for bigger enterprises to develop AI in the future.
- Semi-Cognitive Automation uses both structured and unstructured data, beginning to develop interactions between human knowledge and machine learning, integrating current business processes and leveraging these processes with the use of AI. At this phase, we may begin to identify user personas and gather information about their preferences.
- Eventually, we will arrive at Artificial Intelligence with self-learning capabilities, machine learning, deep learning, and additional tools such as natural language processing, which will be discussed in future sessions.
Mr. Michael then addressed the main confusion that people make when discussing artificial intelligence. Some people assume that RPA is AI, which it is not. In fact, when they are discussing RPA, they are essentially discussing Intelligence Automation or IA. However, after more use cases of intelligence automation we will hopefully be able to evolve into functionality in the full sense of artificial intelligence or AI.
The Ethics of Artificial Intelligence
Mr. Michael then went on to discuss the ethical issues behind AI by first discussing the concept of ‘identity’, defining what a human being will be in the future. Is identity all about entity? Is it the capability for an institution or an individual to be identified as a person that is distinct from others as an independent existence? This is most fundamental discussion about identities. Is identity the set of behaviors that we manifest, not only online, but also to the outside world? In the future systems will be capable of recognizing behaviors and will be able to identify who the person is interacting with the system without them identifying themselves.
Or is identify related to confidential data that is personally identifiable, being collected based on our activities or interaction with the platforms? Could the subject also relate to anonymity, to creating pseudonyms in order to interact with the digital world? These questions give rise to the key discussions that needs to occur moving forward – what is the final version of ‘digital identity’ that we must keep moving forward? What implications do that final version have for us?
Mr. Michael discussed the concept of a “digital twin,” a digital representation of ourselves in the digital world, which is a combination of digital personas and digital relationships. IDC research revealed that every single person in the Asia Pacific Region has 24 digital identifications, on average. This includes digital relationships such as banking and telco relationships, email, and entertainment channels. If there are opportunities for these digital ID relationships to be integrated this would then result in a good representation of ourselves as a human being. This implication builds up into an understanding how AI could potentially engage with us in the digital world. Mr. Michael disclosed the fact that due to the significance of the ‘digital twin’ concept, many large entities such as Google, Apple, Facebook and Amazon have started to integrate ‘digital identities’ into their businesses.
Looking more closely at Thailand, Asia Pacific citizens may have on average 24 digital identities, or this number might be closer to 30. The most common digital relationships and identities that Thai citizens have are higher than average in the social and gaming sectors. On the other hand, Thai citizens have a lower than average number of digital relationships in the banking sector compared to the rest of the region.
Mr. Michael then addressed the risk of AI being abused in Thailand due to the lack of concern for data privacy issues. IDC research also revealed that not only are Thais not concerned about data privacy issues, but that Thai citizens also have the tendency to provide any information needed, no matter how cumbersome the tasks, which could support a more complete formation of a digital identification. Therefore, Mr. Michael gave out a warning that if we do not get the ethics around AI right the personal information provided by Thai citizens may be easily abused.
Mr. Michael then proceeded to make statements on the ethical issues related to how AI should be engaging with us. The first statement made by Mr. Michael was that digital and AI engagement should increasingly be based upon ‘informed consent’. This means that it is the responsibility of the organizations using AI technology to inform their users that they are going to be engaging AI with the users’ digital relationship with the organization, and to obtain the consent of the users. Mr. Michael then described IDC’s Framework for Customer Informed Consent. There are principles that must include the concept of informed consent, particularly the reasons of using data, the flexibility of giving permission to change data privacy, essentially giving the choice of saying yes or no to such use to customers.
The second statement Mr. Michael gave was on the quality of data. It is important to recognize whether data is biased. This is a heavily discussed issue around the facial recognition feature of AI. For example, AI may be biased towards a specific group of people based on the analyses made on the behavior or preferences of people related to a specific type of facial features. Exclusive privileges may be assigned to people based on how they look. This, however, could be a fallacy based on biased data, or based on an inherently biased society. Therefore, it is important for AI adopters to make sure that AI data is not vulnerable and also not manipulated so as to exclude specific people from certain privileges. This especially relates to a potential for a malicious network to potentially trick the system and manipulate its interpretation.
The third statement on how we should use AI related to verifiability and transparency. Mr. Michael stressed the importance of how a customer should be able to change his or her mind, and customers should be informed about why certain decisions were made using their data. At the same time, customers should also be given the opportunity to provide more information if it is going to be more beneficial for them. If we are going to use AI in our organization we would have to look at certain governance frameworks, and how we would be coping with the degradation of data model over time. Closing his session, Mr. Michael concluded that while we are waiting for AI to evolve in its continuum of transformation we need to get the foundations right. We have to be as unbiased and ethical as possible.
Empowering Your Business Transformation with AI by Dhanawat Suthumpun
The next session speaker was Dhanawat Suthumpun, Managing Director of Microsoft Thailand, who had come to share Microsoft’s vision on artificial intelligence; ‘AI is Now’. Mr. Dhanawat opened with an explanatory video of how Microsoft’s AI helps architect the process of bringing history back to life. Engineers explore how AI could help sight-impaired people visualize their surrounding through audio description. Microsoft is realizing the opportunity and creating ‘tomorrow’ today, which underpins Microsoft’s tagline – “Tomorrow is here today.”
Looking back on the progress of AI, in 2016, AI recognition capabilities allowed systems to recognize different objects. In 2017, AI achieved speech recognition at an equal level to human capabilities. In 2018, Microsoft led a study together with Stanford University testing AI capabilities to read and answer 300,000 questions, once again meeting human capabilities up to 89 percent. These tests showed the implications that AI could match four core human capabilities of seeing, hearing, speaking and translating.
Mr. Dhanawat then presented Microsoft’s AI Vision covering core aspects on how human ingenuity could be amplified by intelligent technology:
- Empowering developers to innovate
While some people may be concerned about how AI may replace many jobs the real implication is that AI would help and support people so that they can conduct their job in a more efficient and productive way. AI could be used to empower developers to innovate. In fact, if you are a developer and you are not using AI in your work you are out of the equation.
- Empower organizations to transform industries and empower people to transform society
AI could also be used to transform industries, and even society. Giving some example of how this is possible, Mr. Dhanawat then introduced an application called “Seeing AI” (now available for iOS). This is a Microsoft research project for people with visual impairments, describing the environment and objects or people around them through audio.
Mr. Dhanawat then went on to describe how we could start our digital transformation journey with AI through a digital feedback loop covering 4 areas:
- How to Engage customers
- How to use ai and digital to empower employee
- How to Optimize operation; and
- How to Transform product
Mr. Dhanawat also explained digital feedback loops in relations to these 4 areas. In order to leverage AI capabilities, a continuous data feedback loop must be implemented throughout all areas of interactions, from engaging customers onto engaging employees. With a powerful digital feedback loop, machine learning can be applied to transform products and services to be able to meet more specific needs of the users. For example, a new feature of Microsoft PowerPoint would allow you to be able to give your presentation with subtitles in over 60 languages. This capability of AI could increase the efficiency and productivity of our work on a daily basis. Chatbot is another example of leveraging AI capability into our business.
Mr. Dhanawat then demonstrated another use case of Microsoft AI, applying chatbot use and image recognition to supply electric bills to households by merely submitting their electric meters photos to the bot. The last use case is part of a project Microsoft conducted as an AI Workshop with a business conglomerate, introducing Azure AI tools to their procurement team, assisting in inventory management and product procurement prediction. The workshop required that participants came prepared with one problem statement they would like to work on and real data to support their work. They wrote algorithms to study data from the past three years and predicted patterns of inventory levels. The result has shown that they have the ability to use historical data to predict their inventory for each month, which is ninety-six percent accurate related to actual demand. Microsoft worked closely with the team attending the workshop as a coach in order to ensure that the client can achieve their goals using AI technologies.
Mr. Dhanawat emphasized the importance of data governance on declaring the reasons for data usage, in order to prevent companies from abusing the use of data.
Closing his session, Mr. Dhanawat gave a strong statement on how Microsoft believe that AI could be in anybody’s hands, with increased accessibility and inclusiveness. That means that anyone could engage AI into their business without making a huge investment as was required before. It is no longer a requirement to invest millions of dollars to develop your own AI technology.
Melvin Wong on Leveraging Emotional AI and Robotics to Drive Efficiency
Melvin Wong, PreSales Lead for SAP Innovation Office, SAP Southeast Asia, discussed how emotional AI and robotics could be used to drive efficiency. Opening his session, Mr. Melvin quoted Andrew Ng: “Data is the new oil; Artificial Intelligence is the new electricity.”
While data is the supply of natural resources, AI could create value from those resources and give you more actionable insights on how to make improvements. Mr. Melvin then gave an analogy to AI being the new electricity; where before, if you wanted to move a vehicle with a steam engine, you would need to fuel it with steam and then wait for two hours before it could start moving. But with electricity, you can just start it and move. AI essentially works in the same way. It is not replacing the vehicle, but merely changing the way you operate it.
For SAP, AI is the way to change processing and improve over time.
Mr. Melvin then explained that there are two aspects of AI:
1) General AI, where we are using intelligence that is embedded into a machine and the program to think like a human, which is not yet achievable;
2) Narrow-Version AI, where we use specific forms of AI capabilities, which is essentially where we are right now. But first, we need to understand that there are three layers of AI:
- machine learning,
- deep learning, and
- data science
Today, the majority of where we are is at Machine Learning. Mr. Melvin then went on to demonstrate the case for Machine Learning, where we have a camera installed on our car to detect pixel forms to see movements and recognize patterns of various objects that come into view. The information obtained from one driver’s experience can be shared with the other drivers, and this information can then be leveraged with machine learning to improve their experiences. So the significance of AI could be how we can supplement what we currently cannot do, and augment existing capabilities with machine learning.
Mr. Melvin then went on to describe SAP’s approach to AI, introducing “SAP Leonardo.” SAP not only gives stability to the backend of its customers, but also offers three ways to apply intelligence:
1) Optimize Embedded Intelligence – give ERP system forecasting capabilities
2) Extend Industry Innovation Kits – collect information from the industry and typical use cases, modeling it for use within the industry
3) Transform Open Innovation – enable individual digital transformation journeys
Blockchain, Internet of Things (IoTs), Machine Learning, Robotics Process Automation – these are the new technologies that SAP are trying to adopt across the board, based on the belief that these technologies will facilitate the transformation of industries. SAP’s Leonardo approach aims at supporting its customers along their own journey, providing innovation services at various stages to help them achieve desirable business outcomes. The SAP Leonardo approach runs through five processes as follow:
Starting with exploration, SAP works with customers to reimagine and create new solutions, following the steps of developing a working prototype, testing, iteration, scaling and transformation. The underlying intention of projects is to simplify experiences.
One case Mr. Melvin cited helps improve productivity and efficiency by applying machine learning to understand handwriting for executing commands, fetching reports and information on demand. That was followed by a farming case where a Chicken Farm in China uses AI to detect the health of their chickens.
SAP worked with the client to interpret how chickens speak. The system listens to the noises chickens are making and detects any outliers, which in-turn enables detection any disease in the chicken. The system also uses machine learning to look at the images of how the chickens sleep. From these images, the AI recognizes the pattern of how the chickens group together to identify which chickens are sick.
Mr. Melvin disclosed another use case of AI in a project SAP did with Home Depot to manage and count inventories. Originally, they would need 20 people spending 8 hours to count their inventory each day. Using AI, Home Depot (office supplies providers) would be able to count their inventories more efficiently through image recognition and collaborative robotics (Co-Bots). At the same time, using the drones’ image recognition and analysis, Home Depot could analyze and detect any obstacles blocking certain pathways for moving cargoes and for improving safety measures. Co-Bots could also help reduce manual tasks such as returning merchandize for return policies in a large store setting.
Concluding his session, Mr. Melvin stressed the fact that while AI could be positive in terms of allowing you to achieve wonderful things, the most important question is – does it apply to you? With SAP, the most important thing is trying to understand the customers’ use cases and challenges, and together SAP will help to provide new solutions from the ideation phase to prototyping phase, in order to see whether SAP’s solutions could match the clients’ requirements.
Mr. Melvin closed his session with a bold statement, “AI is always about changing, but are you up for it? Changing for the better involves not only being innovative but also considering sustainability.”