pds-it
['Blog post','no']
Amazon Web Services
Blog

ISO standard for AI: Responsible development through new guidelines

Contents

    An ISO standard for AI does not come as a complete surprise. In a way, the development of AI is still in its infancy and the guidelines are only just being defined. It was only in December 2023 that the EU adopted the first regulations for AI development in the so-called AI Act. The ISO standard for AI now takes things a step further.

    What does the ISO standard mean for AI?

    The International Organization for Standardization ensures global standards for products and services. However, it should be noted that the ISO standard is not a law. Companies voluntarily undertake to comply with international product standards.

    The reason for a standard is primarily the usability of products and services across national borders. This became particularly relevant in the 19th century, when global trade became increasingly important. Products manufactured in Germany should also be able to be used in Spain. The standard ensures this.

    Now also the ISO standard for AI. This standard was adopted by the ISO together with the International Electrotechnical Commission (IEC), which is why the ISO standard number for AI is officially ISO/IEC 42001.

    This is the world's first ISO standard for AI development and use. It is intended to serve as a guide for the development of AI management systems. The newly introduced standard is aimed not only at companies that develop AI management systems, but also at the companies that use these systems. The ISO standard for AI covers topics such as transparency, machine learning and ethical considerations when developing and using AI.

    It is not the first ISO standard for AI in general. This ISO standard for AI exists so far:

    - ISO/IEC 22989 for AI terminology

    - ISO/IEC 23053 for machine learning

    - ISO/IEC 23894 for AI risk management

    AWS directly implements the ISO standard for AI

    Amazon Web Services (AWS) has been helping to develop the ISO standard for AI since 2021 and has also adopted it directly for its own AI systems. AWS had already laid the foundations for the connection to the ISO standard for AI.

    Anthropic, the start-up in which Amazon invested several billion US dollars for the development of AI in 2023, is also committed to the responsible development and use of AI. They should therefore also comply with the ISO standard for AI.

    AWS considers international standards in AI development to be important in order not to lose the trust of customers. "Effective standards help reduce confusion about what AI is and what responsible AI entails, and help the industry focus on reducing potential harm," says Swami Sivasubramanian, the Vice President of Data and Machine Learning at AWS.

    Get to know and deepen AWS AI systems with skill IT

    In our seminars, you will get to know the AI systems in AWS and learn how to use machine learning to solve problems. In our four-day, English-language course The Machine Learning Pipeline on AWS our trainer will prepare you extensively for the AWS Certified Machine Learning (Specialty Level) exercise.

    If you prefer to start small, the one-day course Deep Learning on AWS is something for you. Learn how deep learning works in AWS and how to use deep learning modules with the help of practical exercises.

    Author
    Kia Figge
    As the founder of Textflamme, Kia has been writing for companies from all industries for over 10 years. She has written texts for countless websites and blogs and feels at home in the field of information technology.