Curriculum information

Location of Study

ICT Building, Mahidol University, 999 Phuttamonthon 4 Road, Salaya, Nakhonpathom 73170

Educational Management System

Educational System
Two Semester Credit system. 1 Academic Year consists of 2 Regular Semesters, each with not less than 15 weeks of study.
Summer Semester
None
Typical class hours
1^st Semester: August – December
2^nd Semester: January – May

Graduation Criteria

Duration of graduation: 2 years
Total Credits: No fewer than 36 credits

Graduation requirements:

Plan 1.2 Academic: coursework and thesis
1) Students must complete their courses as stated in the curriculum with a minimum CUM-GPA of 3.00.
2) Propose thesis to the committee appointed by the Faculty of Graduate Studies and to the public and pass oral thesis examination as the final stage
3) The complete or part of the thesis has to be published as a research article, accepted as an innovation, acknowledged as a creative product, or accepted as an academic article that can be searched
4) Other requirements shall follow those that specified by the Faculty of Graduate Studies.

Plan 2 Profession: coursework and independent study

1) Students must complete their courses as stated in the curriculum with a minimum CUM-GPA of 3.00.
2) Students must pass the Comprehensive Examination following Regulations of Mahidol University on Graduate Studies.
3) Propose Independent Study to the committee appointed by the Faculty of Graduate Studies and to the public and pass oral Independent Study examination as the final stage
4) Other requirements shall follow those that specified by the Faculty of Graduate Studies.

Curriculum Structure

* Students who have limitations in computer science fundamental knowledge, can choose to study some courses in the fundamental courses recommended by advisors or program committee. The subjects in the fundamental courses will not be counted in the total credits. The students will be evaluated AU (Audit).

Study Plan

* May be offered in 2nd semester 2 or the summer semester prior to the study

Elective Courses

Students in Plan 1.2 Academic: coursework and thesis can choose to take the following courses not less than 15 credits in additional to the required courses, while students in Plan 2 Profession: coursework and independent study can choose to take the following courses not less than 12 credits in additional to the required courses.

Research Areas

Research Project (for Plan A(A2))
1) Computer Science Theory
2) Software Development and Software Engineering
3) Network and Security
4) Data Management, Data Governance, Data Science and Data Engineering
5) Artificial Intelligence, Machine Learning and Computer Vision
6) Embedded Systems and the Internet of Things

Teaching & Assessment

Program Learning Outcomes

For students following the academic track (Plan 1.2 Academic: coursework and thesis), at the end of the study from the program, students will be able to

PLO1: Apply specialized knowledge of computer science principles and theories to solve computer science-related problems.
PLO2: Conduct research and produce creative computer science solutions with international and publishable quality.
PLO3: Follow the value of ethics and code of conduct in research, academic, and computer science careers.
PLO4: Possess creative, analytical and critical thinking with ability to work independently, responsibly and collaboratively in a team.
PLO5: Communicate computer science research findings to different levels of audiences effectively.

For those pursuing the professional track (Plan 2 Profession: coursework and independent study), at the end of the study from the program, students will be able to

PLO1: Apply knowledge of computer science principles and theories to solve real-world problems.
PLO2: Design and implement methods and techniques in computer science domains to solve real-world problems.
PLO3: Follow the value of ethics and code of conduct in research, academic, and computer science careers.
PLO4: Possess creative, analytical and critical thinking with ability to work independently, responsibly and collaboratively in a team.
PLO5: Communicate practical computer science knowledge and solutions to different levels of audiences effectively.

Future Careers

1. Computer and Information Technology Technical Officer
2. Software and System Developer
3. Information Technology Manager
4. Data Analyst and Data Scientist

Download

• Master of Science Program in Computer Science curriculum

APPLICATION

For prospective students:

Admission Requirements for Plan 1.2 Academic: coursework and thesis – Onsite classroom

1) Applicants should hold a Bachelor’s degree in Computer Science, Computer Engineering, Information Technology, Information and Communication Technology, Electrical Engineering, Mathematics, Physics or related areas.
2) Applicants should have a cumulative GPA of at least 2.5 in a 4.0 GPA system.
3) Applicants should have an English Proficiency Examination score as required by the Faculty of Information and Communication Technology and/or the Faculty of Graduate Studies.
4) Other requirements shall follow those that specified by the Faculty of Graduate Studies
5) Applicants with qualifications other than 2–4 may be considered by the Program Director and the Dean of the Faculty of Graduate Studies.

Admission Requirements for Plan 1.2 Academic: coursework and thesis – Online / Internet-based distance education

1) Applicants should hold a Bachelor’s degree in Computer Science, Computer Engineering, Information Technology, Information and Communication Technology, Electrical Engineering, Mathematics, Physics or related areas
2) Applicants should have a cumulative GPA of at least 2.5 in a 4.0 GPA system.
3) Applicants should have an English Proficiency Examination score as required by the Faculty of Information and Communication Technology and/or the Faculty of Graduate Studies.
4) Other requirements shall follow those that specified by the Faculty of Graduate Studies
5) Applicants with qualifications other than 2 – 4 may be considered by the Program Administrative Committee and the Dean of the Faculty of Graduate Studies.
6) Applicants should have a computer or computing device that meets the minimum requirements of Internet access. A camera, speaker, and microphone must be available on the computer or computing devices for interactive learning experience.

Admission Requirements for Plan 2 Profession: coursework and independent study – Onsite classroom

1) Applicants should hold a Bachelor’s degree and have at least 6 credits of computer related courses or have at least 1 year of work experience related to computing or IT development. The 6 credits of computer-related course may be taken outside of the Bachelor’s degree.
2) Applicants should have a cumulative GPA of at least 2.5 in a 4.0 GPA system
3) Applicants should have an English Proficiency Examination score as required by the Faculty of Information and Communication Technology and/or the Faculty of Graduate Studies.
4) Other requirements shall follow those that specified by the Faculty of Graduate Studies
5) Applicants with qualifications other than 2 – 4 may be considered by the Program Director, and the Dean of the Faculty of Graduate Studies.

Admission Requirements for Plan 2 Profession: coursework and independent study – Online / Internet-based distance education

1) Applicants should hold a Bachelor’s degree with at least 6 credits of computer related courses or have at least 1 year of work experience related to computing or IT development. The 6 credits of computer-related course may be taken outside of the Bachelor’s degree
2) Applicants should have a cumulative GPA of at least 2.5 in a 4.0 GPA system
3) Applicants should have an English Proficiency Examination score as required by the Faculty of Information and Communication Technology and/or the Faculty of Graduate Studies.
4) Other requirements shall follow those that specified by the Faculty of Graduate Studies
5) Applicants with qualifications other than 2 – 4 may be considered by the Program Director, and the Dean of the Faculty of Graduate Studies.
6) Applicants should have a computer or computing device that meets the minimum requirements of Internet access. A camera, speaker, and microphone must be available on the computer or computing devices for interactive learning experience.

Important Dates (Academic Year 2025)

CONTACT US

Asst. Prof. Boonsit Yimwadsana
Faculty of Information and Communication Technology Mahidol University
999 Phuttamonthon 4 road, Salaya, Phuttamonthon, Nakhonpathom 73170 THAILAND
Tel: 66-2-441-0909
Email: boonsit.yim@mahidol.ac.th

Important Dates (Academic Year 2025)

Tuition Fees

* Official Tuition Fees Announcement from Mahidol University

Education fees information

• Education fees information, B.E. 2567 (2025)
• Education fees information, B.E. 2567 (2024)
• Education fees information, B.E. 2564 (2021)

Scholarships

1. Master-level scholarship is available on a case by case basis. Scholarship will be awarded to students who have good academic background, career experience, research experience, computer science skills, English skills, and health. Since the master-level scholarship requires recipients to work as teaching/lab assistant or programmer/system engineer during office hours for a certain period of time during study and after graduation, the recipient must be able to work during day time and after graduation at the Faculty of ICT. In accordance to tuition fee announcement, the scholarship may include one or more of the following items: tuition fee, equipment fee, research fee and monthly allowance.
2. Student exchange scholarship is awarded by the International Relation website. Mahidol University provides support for students seeking opportunity abroad for a period of time. Please see the website of Mahidol University International Relations division for more information (e.g. how to apply, amount of support and qualifications).
3. Partial research scholarship is awarded by the Faculty of Graduate Studies for foreign students. For more information, please visit the website of the Faculty of Graduate Studies .

M.Sc. in Computer Science

M.Sc. in Computer Science program at Mahidol University has produced high quality young professionals to Thailand’s IT industry since 1994.  The program is an international program.  Graduates will have a solid and broad knowledge of computer science and information technology.  Currently, the program focuses on computational intelligence and artificial intelligence.  Most of the information provided here include basic information, tuition fees, and admission information.

International Admission Schedule

Academic Year 2023 Enrollment

Second Semester Round

Academic Year 2024 (Tentative information)

Second Semester Round

Reference: https://graduate.mahidol.ac.th/inter/prospective-students/?topic=schedule

Admission Procedure

Apply Officially This active recruitment session is an interactive Q&A session. To be considered for the program admission, please apply officially online at Apply Online: https://graduate.mahidol.ac.th/admission-inter/login.php Please note that the program admission committee will not be able to consider candidate’s eligibility without the candidate’s complete application form and admission fee payment. Selecting Academic Program from https://graduate.mahidol.ac.th/inter/prospective-students/?topic=curriculumOpened&degree=m to apply: A candidate can select the computer program from

• 6901MG00 MASTER OF SCIENCE PROGRAM IN COMPUTER SCIENCE (INTERNATIONAL PROGRAM) for regular program (teaching during daytime of weekdays – preferred for international student applicants)
• 6901MS00 MASTER OF SCIENCE PROGRAM IN COMPUTER SCIENCE (INTERNATIONAL PROGRAM) (SPECIAL PROGRAM) for special program (teaching in the evening of weekdays and daytime of weekends – preferred for international student applicants who are working daytime as fulltime employee)

Plan of Study

Plan A: thesis

• Student must complete a thesis defense and publish in an international journal or conference proceeding to graduate.
• A thesis is an original research project that are conducted using scientific method and research methodology. A thesis at the Master’s level do not require a rigorous and strong research scope like a thesis at the Ph.D.’s level.  A thesis at the Master’s level could be a small improvement of an existing research.

Plan of study for Plan A

Plan B: thematic paper

• Student must pass all courses, comprehensive examination and complete a thematic paper defense.
• A thematic paper is an independent study of a specific area of academic interest. Usually, a thematic paper is related to a validation of an existing research or a real-world problem-solving task using the knowledge and skills from the program.

Please note that in some classes, students in regular program and special program may study in the same class due to teaching resource limitation.

Plan of Study and Requirements for Admission and Graduation

Plan A: Admission requirements

• Applicants should hold a Bachelor’s degree from an institute accredited by the Office of the Permanent Secretary, Ministry of Higher Education, Science, Research and Innovation, in either one of the following categories:
  • A degree in computer science, computer engineering, information technology, information and communication technology, electrical engineering, mathematics, or physics.
  • A degree in another related field with at least 12 credits of computer related courses, and having at least 1 year of work experience in computing or IT development.
• Applicants should have a cumulative GPA of not less than 2.5
• Applicants should have an English Proficiency Examination score as required by the Faculty of Graduate Studies.*

Graduation Requirements

• Total time of study should not exceed the study plan.
• Students must complete courses as stated in the curriculum. At least 24 credits excluding thesis (12 credits) for 36 credits in total, with a minimum CUM-GPA of 3.00.
• Students must meet the English Competence Standard of Graduate Students, Mahidol University as defined by the Faculty of Graduate Studies, Mahidol University.
• Students must participate in skill development activities by the Faculty of Graduate Studies, Mahidol University.
• Students must submit theses and pass the thesis defence in accordance with the Regulations of Mahidol University on Graduate Studies and the oral thesis defense must be open to public.
• Theses are required to be published in an international academic journal or proceedings that is listed by the Faculty of Graduate Studies, Mahidol University.

Plan BAdmission requirements

• Applicants should hold a Bachelor’s degree with at least 6 credits of computer related courses from an institute accredited by the Office of the Permanent Secretary, Ministry of Higher Education, Science, Research and Innovation and have at least 2 years of work experience in computing or IT development.
• Applicants should have a cumulative GPA of not less than 2.5
• Applicants should have an English Proficiency Examination score as required by the Faculty of Graduate Studies.*

Graduation Requirements

• Total time of study should not exceed the study plan.
• Students must complete courses as stated in the curriculum at least 30 credits excluding the thematic paper (6 credits) for 36 credits in total, with a minimum CUM-GPA of 3.00.
• Students must meet the English Competence Standard of Graduate Students, Mahidol University as defined by the Faculty of Graduate Studies, Mahidol University.
• Students must participate in skill development activities by the Faculty of Graduate Studies, Mahidol University.
• Students must pass the comprehensive examination following Regulations of Mahidol University on Graduate Studies.
• Student must propose and complete a thematic paper and pass the oral thematic paper examination required for graduation according to regulations of Faculty of Graduate Studies, Mahidol University and the oral thematic paper examination must be open to public.
• The thematic paper or a part of thematic paper must be published and searchable.

*For academic year 2024, students must pass the English admission requirements as follows in order to be officially admitted into the program.

For more information about the regulations and MU GRAD Plus test registration: visit the website of the Faculty of Graduate Studies language center https://graduate.mahidol.ac.th/thai/current-students/?g=6

• IELTS, MU GRAD Plus requires at least 3 weeks to register for the exam.
• TOEFL-iBT requires at least 3 months to register for the exam.
• MU ELT is only available to Mahidol University students and staff.
• Students who do not have these scores by admission time will not be admitted into the program.
• The test scores is valid for 2 years after the test date.
• Exemption:
  • Citizen of English native speaker countries.
  • Graduate from accredited higher-education institutions of English native speaker countries
  • English speaker countries: United Kingdom, United States, Canada, Australia, New Zealand, Ireland, Singapore, and South Africa.
• There are two English requirements: admission requirements and graduation requirements. It is best to pass the English graduation requirements since the admission.

Tuition Fees

Academic Year 2023

Academic Year 2024

Note: extra fees applied if students cannot graduate within 2 years.

Study Location

Courses are provided primarily on-site at the Faculty of Information and Communication Technology, Salaya campus, Nakhonpathom.  It takes around 45 minutes to travel between Salaya campus and Bangkok city center. Online access: Some courses provide online access as supplementary to on-site class for class review purpose.  Students are required to study on-site.

Class time

Special program: All classes are offered in the evening of weekdays and weekends (outside regular office hours)

Regular program: All classes are offered during the daytime of weekdays (regular office hours 9:00 – 17:00)

Scholarship

There is currently no scholarship for Thai and international candidates.  However, Mahidol University do provide allowance scholarships for international students.  For more information, please visit https://op.mahidol.ac.th/ir/mu-scholarships/

Remarks

The information provided in this document is subjected to change without notice due to the change in university and faculty policies over time.  For most updated version of this document, please contact the program director or the program staff.

Contacts

Program director: Asst. Prof. Boonsit Yimwadsana (boonsit.yim@mahidol.ac.th)

Program coordinators: Thunyathorn Suttijaroen (thunyathorn.sut@mahidol.ac.th), Buntida Suvacharakulton (buntida.suv@mahidol.ac.th)

Artificial Intelligence

Article by Boonsit Yimwadsana (last updated 30 September 2024)

What is AI today?

Artificial Intelligence (AI) is broadly defined as a field of study and type of technology characterized by the development and use of machines that are capable of performing tasks that usually would have required human intelligence [1].  Today AI has already penetrated many industries and societies in such a way similar to how the Internet has transformed our lives (both business and leisure).  In 2025, almost everyone has heard of ChatGPT, an intelligent software that can answer any question.  It is more convenient for people to find an answer than having to go through search results provided by popular search engines.  AI today has the ability to comprehend our questions, predict the answers to the questions, and answers questions beautifully in our languages without requiring human intervention.  People today perceive AI as a magic software tool that can find solution to anything.  Thanks to its vast knowledge and ability to speak human language, AI can definitely transform how to conduct our businesses and daily activities.

AI’s greatest potential is in the applications of human imitation.  When scientists invented computer machines, we expected that the computers will handle our routine tasks and finally replace us from workforce.  However, those routine jobs were very dull, and uninteresting.  They do not require any forecasting or decision making because the forecasting and decision-making processes were carried out by human.   Artificial Intelligence today has the ability to breakthrough this limitation.  Human complex tasks of forecasting and decision-making today have already been developed and implemented as software algorithms and codes which run on powerful computer hardware not available in the past.  Thanks to this development, one of the most useful human tasks, the ability to find knowledge, can be performed by AI.  We can now expect that more human tasks would one day do not require human to perform forecasting and making decision for the machine, and the machine can be left alone to make its own decision without human as an instructor or supervisor.

Misconception about AI

Even though AI is defined as artificial intelligence, the AI technology that we are using today is not yet intelligence (as of 2025).  AI tools perceived today as intelligent machine require man-made algorithms that can perform prediction.  The AI tools itself cannot code or modify its own algorithms that mimic human intelligence.  However, with the introduction of the celebrated large AI models which learn useful relationship patterns from examples in a large amount of data, the AI tool can use the relationship pattern to accurately predict and execute a desired outcome. But it would be a mistake to refer this prediction ability to human intelligence because the AI tools still require human intelligence to learn how to capture relationship patterns in the data.

Core Principles of Artificial Intelligence

A true artificial intelligence system must be able to perform tasks that typically require human intelligence. These tasks include perception (collecting and interpreting sensory input), reasoning (drawing conclusions from rules or facts from the input), learning (finding relationship patterns from the input data and conclusions), planning and decision-making (deciding a goal and optimizing sequences of actions toward the goal, language understanding and processing (understanding and generating human languages), problem-solving (understanding human problems and offer the best solution to the problems from several possible solutions), autonomy (operating independently in dynamic environments) [2].

Evolution of the field of Artificial Intelligence

In order to create a system that is capable of performing the core principles of artificial intelligence, computer scientists and engineers have developed artificial intelligence through several generations which are

• The Symbolic Era (1950s – 1980s): Early artificial intelligence software relied on tailored-made rules and logic systems (e.g., expert systems) for specific tasks. The rule-based systems require man-made rules and logic to function.  The main challenge of these systems is that the systems lacked adaptability and flexibility to cope with real-world complexity and incomplete information.
• The Statistical and Machine Learning Era (1980s–2010s): In this era, artificial intelligence systems learn from the statistics of data to suggest a set of actions rather than relying on tailored-made rules and logic systems. The main challenge of these systems is that the systems require a large amount of quality data and high-performance computers to accurately understand and use the statistics of data within a reasonable amount of time.
• The Deep Learning and Generative AI Era (2012–present): In this era, artificial intelligence systems learn relationship patterns through mathematics, in particular linear algebra, rather than statistical models. Neural networks with many layers (also called deep learning) and  transformers (data-encoding), and attention (key-information extraction) powered state-of-the-art results in image recognition, speech, and translation.  The reverse engineering of deep-learning creates tools which can generate data from context (previously referred to as the understanding or the conclusion of the data).  The main challenge of these systems is that the models are often biased and contain hallucination (confabulated outputs).  Moreover, the models trained on the data often lack explainability.  Due to the vast benefits of these systems, the hardware and data required for training these models have become highly in demand.  This drives the cost of developing these systems exorbitantly high in addition to the cost of environment which has to bear the pollution from the heat and carbon emission generated from the high-performance computing hardware required for running artificial intelligence systems.

What’s next?

Given the challenge of the Deep Learning and Generative AI Era, the field of artificial intelligence is heading towards the safe, ethical, explainable, affordable, and sustainable use of artificial intelligence.   In addition, the usability of artificial intelligence will expand from conventional information and knowledge finding towards task automation in the aspect of independent decision-making and actions.  It is widely expected that future deep learning and generative AI models will incorporate a large database of action and behavioral data so that machines can automate actionable tasks without human intervention.  Possibly, machines would be able to evaluate its own decision and perhaps find reasons to support the decisions made by the machines.  Furthermore, when AI systems have become widespread, the ethical use of artificial intelligence systems must be handled by authorities and societies.  There are already examples of the malicious and unethical use of artificial intelligence tools.  The detection of work produced by artificial intelligence tools and laws related to the intellectual property ownership of artificial intelligence tools must be established in order to ensure that artificial intelligence is used ethically to support the development of the society.

References:

• [1] World Economic Forum: “What is artificial intelligence—and what is it not?” https://www.weforum.org/stories/2023/03/what-is-artificial-intelligence-and-what-is-it-not-ai-machine-learning/, 8 Mar 2023
• [2] Russell, S. J., & Norvig, P., “Artificial Intelligence: A Modern Approach”, (4th ed.). Pearson, 2020

Machine Learning

Article by Boonsit Yimwadsana (last updated: 25 September 2024)

In 1959, MIT engineer Arthur Samuel [1] described machine learning as a “Field of study that gives computers the ability to learn without being explicitly programmed.”  Machine Learning is a sub-area of artificial intelligence.  The main goal of artificial intelligence is to imitate human intelligence.  In the first era, artificial intelligence is implemented as rules designed and created by human.  A major difficulty in using artificial intelligence in the first era is that human intelligence is usually complex and consist of many conditions and exceptions.  Artificial Intelligence in the first era is quite inflexible and unable to be applied smoothly in real-world settings.  In the second era, instead of human having to come up with complex rules, machine learning algorithms are invented to learn and estimate human decision-making process from input and output data.  There are four types of machine learning algorithms:

1. Supervised learning: In supervised learning, the machine is taught by examples. The operator provides the machine learning algorithm with a known dataset that includes desired inputs and outputs, and the algorithm must find a method to determine how to arrive at those inputs and outputs. This method requires a large number of examples to be able to accurately predict outputs from inputs.  Conventionally, a foundation relationship model such as linear regression, Bayes’ statistics, Decision Tree, Random Forest, Support Vector Machine, and neural networks, must be first estimated prior to the actual learning which provides a precise relationship model.  There are 2 major types of machine learning tasks associated with supervised learning: prediction and classification.
2. Unsupervised learning: unsupervised machine learning algorithms study data to identify patterns. There is no answer key or human operator to provide instructions or examples. Instead, the algorithms determine the correlations and relationships by analyzing available data based on different foundation relationship models. In an unsupervised learning process, the machine learning algorithm is left to interpret large data sets in order to find relationship or knowledge of that data accordingly. The algorithm tries to organize the data in different way in order to meaningfully describe its structure.  Typically, grouping or segmenting the data into clusters or arranging it in a way that looks more organized are common unsupervised learning tasks..
3. Reinforcement learning: Reinforcement learning algorithms focuse on regimented learning processes, where the algorithms are provided with a set of actions or rules, parameters and end values. By defining the rules, the machine learning algorithms then try to explore different actions, options and possibilities, monitoring and evaluating each result to determine which set of actions, options and possibilities is optimal. The optimization method is often based on classical numerical methods.  Reinforcement learning algorithms allow the algorithms to learn the optimized relationship model that can meaningfully describe the data through the process of trial and error.  The algorithms learn from past experiences and adapt their approach or the set of actions, options and possibilities to achieve the best possible result using optimization methods.
4. Semi-supervised learning: Semi-supervised learning algorithms are similar to supervised learning algorithms, but instead of using both labelled and unlabelled data in examples. Labelled data is essentially the information that has meaningful tags so that the algorithm can understand the data using only labelled data. The unlabelled data lacks the information of unlabelled data.  By using this combination, machine learning algorithms can learn to label unlabelled data.  This type of semi-supervised learning is often used to predict anomaly in the data.

References:

1. L. Samuel, “Some Studies in Machine Learning Using the Game of Checkers,” in IBM Journal of Research and Development, vol. 3, no. 3, pp. 210-229, July 1959

AI Agents vs. Agentic AI

Article by Boonsit Yimwadsana (last updated: 15 November 2024)

AI Agents

Traditional AI systems (1^st era) focus on the applications of computing rules and logics that process inputs to create outputs according to human intelligence.  Recent AI systems (2^nd era) allow machines to learn relationship models that can imitate human intelligent processes from a large amount of example data.  Present and near future developers can use the power of AI systems as a service (usually a networked or web services) via the service’s APIs.  Each individual service performs its own specific task intelligently based on the developer’s needs.  By tying different services together into a chain of services, an AI process or workflow is created.  Each individual service is called an AI agent.  AI agents have revolutionized the field of computational intelligence.  Rules and logics do not have to be constantly adjusted manually.  Instead, machine learning systems in the name of AI agents learn the rules, relationship, and logics with high precision for prediction, classification and clustering from a large amount of example data.

When multiple different AI agents work together, their capabilities and benefits can increase exponentially.  Instead of focusing on just one part of a process, a team of AI agents (or, an agentic AI system) can collaborate to handle complex workflows and achieve more advanced and complex goals. For instance, in customer service, one AI agent might process language, another might search knowledge bases, and a third could handle ticket routing—all working together to solve customer issues efficiently. The AI agents are so widely used today in several advanced applications that required complex process workflows such as biometric-based authentication systems used by major financial corporation, cancer detection from x-ray images used by advanced medical institutions, and customer segmentation used by marketing research firms.

Agentic AI

Agentic AI is a broad concept describing the sophisticated setup that integrates multiple AI agents together, allowing multiple agents to work together to achieve the goals of an organization.  Agentic AI is different from an AI agent workflow in the aspect of how agents work together.  The Agentic AI concept requires AI agents to work together autonomously and automatically.  Typical AI agents require human intervention to create a complex workflow.  Each service provided by an agent may be manually activated by an individual resulting in a manual workflow instead of an automated workflow.  Organizations can use Agentic AI systems to automate streamline AI agents responsible to different processes to reduce processing steps, time and cost to achieve the organizations’ goals.

Generative AI

Article by Boonsit Yimwadsana (last updated: 15 February 2025)

Generative AI refers to a category of AI algorithms that generate new outputs based on the data they have been trained on. It uses a type of deep learning called generative adversarial networks and has a wide range of applications, including creating images, text and audio.  Generative AI works by learning patterns from large datasets and then generating novel content that resembles the learned patterns. Think of it as an AI “artist” that can create something new based on examples it has seen.  Traditionally, we believe that AI cannot perform creative tasks and human jobs are still safe from being replaced by AI.  The rise of Generative AI change the whole picture of job market.  By using neural network technology to identify the patterns and structures within existing data, generative AI creates new and original content. The technology can be a powerful tool for streamlining the workflow of creatives, engineers, researchers, scientists, and others—across all industries.

Generative AI has become explosively popular thanks to the release of an intelligent chatbot, called ChatGPT, created by OpenAI in 2023.  The impact of ChatGPT to the world can be compared to the same level of the impact of Google search engine to the world almost 30 years ago.  According to a recent McKinsey report, generative AI is now being used by 79% of workers and 22% use it regularly to complete their day-to-day work tasks.

Generative artificial intelligence (GenAI) could significantly boost productivity while reshaping many jobs. By aligning strategic goals with the needs of their people, organizations can create an environment where GenAI improves job quality, productivity, and helps employees take on more meaningful and impactful work.  While there are concerns about the impact of AI on the job market, there are also potential benefits such as freeing up time for humans to focus on more creative and value-adding work.