Creating a strong foundation for trusted AI applications

Gorodenkoff/Shutterstock.com

Effective and efficient life cycle processes of any software system can significantly improve quality, cost and performance. Artificial intelligence (AI) systems have many processes that are specific to them, which is why a new international standard has just been published.

AI systems such as machine learning, natural language processing, speech recognition and information retrieval use a combination of traditional software elements such as code and databases as well as those specific to AI. In machine learning, for example, part of the system is defined by training it on data as opposed to directly inputting code. The data also changes over time so retraining of the system may be needed.

International standards for these processes can provide a strong platform on which developers and users can communicate, thus improving understanding and performance and reducing costs.

ISO/IEC 5338 is an internationally agreed way of describing an AI system life cycle that takes specific AI characteristics into account. It will help to improve understanding of the functions and performance of AI systems during development and promote mutual understanding between developers and users of AI systems.

The standard defines a set of processes and associated terminology for describing the life cycle of AI systems. It is based on existing standards, ISO/IEC/IEE 15288 and ISO/IEC/IEEE 12207, for system and software life cycle processes, with adjustments made to support the definition, control and improvement of AI systems. ISO/IEC 5388 covers everything from key concepts to risk management, stakeholder needs, verification, maintenance and disposal.

All of these standards can be found in our e-shop.