Why artificial intelligence needs standards

Artificial intelligence (AI) is an enabling horizontal technology. Many industry experts and analysts believe that the rapid growth in AI will enable the next digital transformation. While the field of AI is not new, an international standards committee looking at the entire AI ecosystem is a recent development.

To better understand why standards are important for AI, it is helpful to identify the diverse stakeholders involved, which include research, academia, industry, practitioners, policy makers, ethics advocates and more.

Moreover, there are many different application areas of AI technology, for example, consumer, industrial and commercial amongst others. Regarding the industrial sector, IT is becoming more pervasive in the various industry verticals, including manufacturing, healthcare, robotics, and financial.

Standards are required to enable mass deployment and adoption of AI in these fields. As a basis, it will be important to have common terminology for use by all stakeholders, which will enable clear communication and sound decision making. Gathering use cases, their requirements and best practices for application of the technology will guide technology development. Like other transformational IT technologies, AI will be widespread and addressing issues of trustworthiness from the get-go is needed. Finally, looking at the core of AI, standardization of algorithms and computational techniques will allow a higher level of adoption, use and interoperability.

The role of ISO/IEC JTC1 in AI standards

ISO/IEC JTC 1/SC 42 is the first international standards committee that is looking at the entire AI ecosystem. In the creation of JTC 1/SC 42, JTC 1 scoped SC 42 to be a systems integration entity to work with other ISO, IEC and JTC 1 committees looking at AI applications. As stated in the scope, JTC 1/SC 42 will:

  • Serve as the focus and proponent for the JTC 1 AI standardization programme
  • Provide guidance to JTC 1, IEC, and ISO committees developing AI applications

Work programme

SC 42 has set up the following groups which will cover specific aspects:

Foundational standards working group (WG1)

With such AI diverse stakeholders, there is a need for foundational standards which can provide a framework and common vocabulary. This will enable the stakeholders to talk the same language and set the stage for how they and technology providers/users interact with one another.

This group is developing AI concepts and terminology ISO/IEC AWI 22989 and a framework for AI systems using machine learning ISO/IEC AWI 23053.

Use cases and applications study group (SG 3)

Use cases are the currency by which standards development organizations collaborate with each other. As both the focal point of AI’s role as an enabling horizontal technology and in its role as an AI systems integration entity committee tasked with providing guidance to ISO, IEC and JTC 1 committees looking application areas, it is essential for SC 42 to collaborate with other committees and bring in their use cases. By way of example, use cases provided by other committees looking at different vertical application areas can allow for the distillation of technical requirements that the SC 42 committee can take into account as it drafts its standards, technical reports and best practices.

Computational approaches and characteristics of artificial intelligence systems study group (SG 1)

Computational approaches and algorithmic techniques empower the insights provided by the AI engines. ICT advances, specifically computational power, distributed computing methods and software capability techniques amongst others, allow for what once was science fiction to become science faction. Standardization and best practices are essential to allow for innovation to occur over open standards.

Trustworthiness study group (SG 2)

AI is set to join other ICT technologies that have become ubiquitous in our lives. Leading industry experts in this field believe that an essential aspect to wide-spread adoption is the need to address trustworthiness issues, including robustness, resiliency, reliability, accuracy, safety, security, privacy from the get-go with standards and best practices.

Big data

The JTC 1 big data programme now comes under SC 42 and has two foundational projects for the overview and vocabulary and a big data reference architecture (BDRA).

SC 42 will also cover certain societal concerns in its programme.