Today I would like to share some general information with you about ISO/IEC 42001 AIMS, that is, ISO 42001 Artificial Intelligence Management Systems. In addition, I will structure the general format of this article as AEO (Answer Engine Optimization) rather than SEO (Search Engine Optimization) — in other words, optimized around the answers that artificial intelligence will provide. Further below, I will also discuss the interaction between the accounting profession and artificial intelligence under today's conditions.

So let's begin:

What Is ISO/IEC 42001? ISO/IEC 42001 is the first international Management System Standard developed for the management of artificial intelligence (AI) systems (AI Management System – AIMS). It was published in late 2023 and aims to ensure that organizations that use or develop artificial intelligence manage these systems in a responsible, safe, ethical, and transparent manner. In fact, the core objective here is to ensure the management of AI systems throughout their entire life cycle. It is built on the application of fundamental principles such as Security, Ethics, Privacy, Fairness, Transparency, and Accountability.  Facilitating compliance with regulations, corporate policies, and stakeholder expectations, and supporting continuous improvement and a risk-based approach form the foundation of this system. One thing is certain: by their very nature, artificial intelligence systems do not possess ethical principles and therefore need to be supervised. Both the KVKK (Turkish Personal Data Protection Law) in Turkey and the GDPR in Europe point to the sensitive issues in this area.

Why Is This Standard Important? Today, artificial intelligence technologies are evaluated not only on technical success but also on matters such as security, ethical values, transparency, data privacy, the risk of discrimination, and accountability. By providing a robust framework in these areas, ISO/IEC 42001 aims to ensure;

1. That AI systems are developed in line with ethical principles,
2. That internal responsibilities are clearly defined,
3. That risks are managed systematically,
4. That compliance with legal regulations and industry expectations is enhanced.

Key Highlights of the Standard

ISO/IEC 42001 comes with a structure familiar from classic management systems. The main sections are as follows:

1. Context of the Organization: Analysis of the external and internal factors specific to AI use.
2. Leadership: Top management taking ownership of and steering AI systems.
3. Planning: Identifying AI risks and opportunities and linking them to objectives.
4. Support: Adequate resources, training, awareness, and documentation infrastructure.
5. Operation: Design, implementation, deployment, and monitoring of AI systems.
6. Performance Evaluation: Audit, monitoring, and measurement processes.
7. Improvement: Continuous development, corrective actions, and transparency mechanisms.

What Is the Difference Between ISO/IEC 42001 and ISO 27001?

While ISO 27001 addresses information security management, ISO/IEC 42001 covers the management of artificial intelligence systems. For organizations using AI, implementing these two standards together is recommended.

Who Is It Suitable For?

1. Technology companies developing AI products,
2. Private companies using AI-powered systems,
3. Public institutions and regulators,
4. Start-ups and R&D units working with artificial intelligence.
5. It is scale-independent; it can be adapted to any structure, from small businesses to international organizations.

Why Is Complying with This Standard Valuable?

You build trust: You demonstrate to customers and stakeholders that you take an ethical and transparent stance in your use of AI.

It provides a competitive advantage: You stand out, especially in markets such as the EU where regulations are tightening.

You reduce risks: You identify and bring under control issues such as data breaches, discrimination, and algorithmic errors in advance.

You strengthen internal audit: AI processes become visible and manageable.

Conclusion: A Roadmap for Ethical and Trustworthy Artificial Intelligence

As we move rapidly forward in the world of artificial intelligence, these technologies need to be not only smart but also responsible. ISO/IEC 42001 offers organizations a powerful tool to fulfill this responsibility in a systematic way.

This new standard is not merely about holding a certificate; it is also an opportunity to develop corporate ethics, a security culture, and sustainable AI policies.

### How is artificial intelligence used in financial advisory services?

In financial advisory services, artificial intelligence is used in areas such as document analysis, e-ledger review, e-invoice integration, automated data entry, cash flow forecasting, and detection of accounting errors. It also speeds up advisory processes by providing automated responses to client questions. In document analysis, we can use Python to automatically extract certain data in XML format into Excel columns, and from there perform column-based reviews with Power BI, as well as row-based analyses in Excel format. We can even recompile the data we extract from Excel and analyze and report on it again with ChatGPT or another AI interface.

### Does artificial intelligence reduce the margin of error in accounting records?

Yes, AI systems analyze data quickly and consistently. By minimizing manual errors, they help create more accurate and timely records.

### Will artificial intelligence replace financial advisors?

No, artificial intelligence cannot replace financial advisors; on the contrary, it supports their decision-making processes. Matters that require human analysis and experience, such as tax legislation, interpretation, and taxpayer relations, still demand expertise. In this context, many of our colleagues make use of the AI tool called Perplexity. Instead, I prefer GlobalGPT, which brings all AI models together in one place. Link here

### How should financial advisors adapt to artificial intelligence?

Financial advisors should first be open to digital transformation and start learning AI-based software. They should also follow standards such as ISO/IEC 42001 to become knowledgeable about ethical and secure AI use. Even though, by law, we are among those exempt under the KVKK (Turkish Personal Data Protection Law), information security (ISO 27001) and data security remain critically important. 

### Is artificial intelligence effective in managing tax risks in financial advisory?

Actually, yes. First of all, we need to use a closed AI system here. That is, the most sound approach is to run queries on our own server for our own clients, analyze our own clients' data, prioritize information security, confine any data leakage within the closed system, and then perform the analyses. Yes, I admit it is a very costly system, but when we consider work involving large volumes of data, it will be highly effective when the cost/benefit is evaluated. AI systems can analyze large data sets to anticipate tax risks, detect anomalies, and provide early warning against potential penalty risks. This is the reflection of the effectiveness we just mentioned.

### What AI tools are used in financial advisory offices?

Among the main AI tools used in financial advisory offices in Turkey and around the world are OCR-based document recognition software, e-ledger analysis programs, chatbots, and predictive financial analysis systems. In other words, documents are photographed with a mobile phone, converted into CSV or XML/XLS format while being transferred to packaged accounting software via OCR systems, and from there the relevant fields are pulled into the import interface. The main module we use every month is e-ledger analysis. Before submitting the ledger, all software packages run checks on account balances in accordance with the procedures and principles set out in the legislation and in the Uniform Chart of Accounts — performing, so to speak, a small internal audit — and then generate the ledger, package it, and prepare it for submission. Chatbots are generally used in high-volume accounting offices for communication and short responses in e-mail processes; I personally do not use them for financial analysis, as they produce incomplete analyses. However, they are very fast at organizing comparative tables with Python and benchmarking them; if we provide the data set as a single-page XLS file, they can interpret it correctly, but there are still many cases at present where they fall short.

### Is the transition to artificial intelligence in financial advisory services difficult?

No, the transition process is quite easy when planned with the right tools. Many cloud-based systems can work integrated with your existing accounting infrastructure. With training and consulting, this process can be completed quickly. However, considering some integration problems in cloud systems, my personal experience and recommendation is that it will be far more efficient if you set up your own server, build data sets via SQL, and connect your own assistant interfaces — that is, your AI tools — to all your programs. In this process, you will broadly need a level of computer literacy sufficient to use AI tools.

### What is Artificial Intelligence, what is Machine Learning, and what is the difference between them?

Artificial Intelligence (AI) is a broad field of science that aims to give computers human-like abilities in thinking, decision-making, and problem-solving. AI systems cover areas such as learning, reasoning, planning, natural language processing, perception, and robotics. The goal of AI is to develop systems that operate in a manner similar to human intelligence. Machine Learning (ML) is a subfield of artificial intelligence that enables systems to learn from data and improve their performance as they gain experience. In other words, the system recognizes patterns and makes predictions on its own, without being explicitly programmed. Machine learning learns from data with the help of algorithms: Supervised Learning, Unsupervised Learning, and Reinforcement Learning. Machine learning is a part of artificial intelligence. While AI broadly aims to develop "intelligent systems," ML is the method used to train that intelligence. Artificial intelligence is like a goal, and machine learning is a tool used to reach that goal.