Insights 20 March 2024

Bluebik sees ISO/IEC 42001 guarantees ethical and risk-free use of AI and recommends 4 ways for businesses to effectively use AI

Bluebik Group Public Company Limited (BBIK), a leading consultancy on end-to-end digital transformation, is aware that businesses are facing 3 major challenges concerning their development and application of artificial intelligence. Thus, it recommends the AI standard, ISO/IEC 42001, for businesses to reduce risks and ensure ethics and transparency in their AI use. Bluebik also shares 4 ways to help businesses develop their AI with strategies.

Phiphat Prapapanpong, Director of Advanced Insights at Bluebik Group Public Company Limited (BBIK), said AI has become a key tool for growth. It not only enhances efficiency but also supports the innovations that create business opportunities. Any organizations that want to increase competitive edge must study and apply AI. However, many business organizations are facing 3 huge challenges in their AI development. They are ethical challenge, AI use challenge and technical challenge.

  The ethical challenge concerns especially biased AI models. In risk scoring, for example, if distorted data are used to train models, some groups of people may not have access to some kinds of financial services. Another issue under debate is the privacy and security of personal data in relation to the use of data to train models.

The second challenge centers on AI development and its real-world application by businesses. They may lack the strategies that clearly identify the areas of operations where AI is applied to enhance efficiency. They may not have the use cases that suit their business operations. In addition, their personnel may not have knowledge or understanding about AI, or personnel’s mindset is based on the belief that AI is a threat rather than assistance. Also, businesses may not evaluate their AI application.

The third challenge, the technical one, is about the lack of data readiness. For example, there are no quality data for use. Data are incomplete or inadequate, so they cannot be used to train models. Personnel may lack technical knowledge that is essential for the development of analytical models. In addition, there are difficulties in integrating outcomes from models into business operations.

Mr. Phiphat explained that given the challenges, there must be frameworks for AI development and application. A solution that has been widely discussed internationally is AI standard for organizations, known as “ISO/IEC 42001”. It is the standard for the management of AI systems. It guides organizations on risk management and opportunities related to AI application. The standard covers ethics, transparency and work processes. With the AI standard, organizations can enhance their development potential and AI application in 3 areas as follows.

1) The development and application of AI are more secure when it is used for the development of products and services. This is because, for example, there are methods to reduce the risks of data leakage when data are used to train AI models. As well, there are measures to cope with cyberattacks and prevent data losses.

2) Discrimination is lessened and fairness is promoted when it comes to the use of data supplied by AI. Also, the risks of personal data breach are reduced.

3) Organizations have transparency in their AI systems. There are processes for data disclosure and algorithms for the training of AI models. Also, there are frameworks and regulations on the duties and responsibilities of people concerned.

Bluebik, as a leading consultancy on end-to-end digital transformation, sees that organizations can set standard for their systematic development of AI. Bluebik provides advice on the efficient application of AI and the advice consists of 4 main ways.

1) Define AI Strategy – There must be the goals and objectives of AI application. This covers data strategy, use case generation and advice on the legality of AI application by organizations based on the Personal Data Protection Act.

2) Determine Risks & Govern AI – This concerns the assessment of risks related to AI application. There must be guidelines on data governance for the supervision of data use and risk reduction. As well, there must be standard for data quality.

3) Develop Advanced Insights Management – There must be the structures and data management system that support the use of data to develop AI models. This refers to data architecture and data pipelines to guarantee the efficient use of data. In addition, there must be data platforms to facilitate long-term data analyses and data use.

4) Democratize AI Adoption – Organizations should be advised on how their internal management can offer access to new technologies and their application. Organizations should train their personnel and assess their data and AI capability.

“Although AI standard is not obligatory, business organizations should have it because it means good governance in organizations, especially the ones that develop and apply AI systems. As well, AI standard cuts the costs of AI system management, reduces the risks of evaluation without AI and minimizes the risks of data breach and data leakage. AI standard and AI application strategies are crucial factors in the creation of AI-powered opportunities and long-term business growth,” Mr. Phiphat concluded.