ISO set to release an AI management system standard this year

The International Organisation for Standardisation (ISO) is set to release a new standard later this year, the ISO/FDIS 42001, which focuses on AI management systems.

On the ISO site, it is currently "under development". This standard will ask for every AI system that an organisation has to have three governance checkpoints.

This standard will be similar to the cybersecurity standard of ISO 27001, which changed how partnerships and suppliers will be decided. 

At the 2023 TechLeaders Conference, Stela Solar, director at the CSIRO National Artificial Intelligence Centre told Digital Nation the importance of AI management systems and how this will impact business leaders. 

“It is asking for three governance checkpoints of every AI system. It is asking for pre-design, pre-deployment, and post-deployment. Then a robust governance model to have this ongoing governance approach,” Solar explained.

“It's in essence going to be identifying which organisations are more mature with AI governance. By default, organisations who embrace those standards will be demonstrating that they're more mature in their AI practice and governance, those who are not may be seen as higher risk.”

Solar said she anticipates shifts in the supply chain accordingly as organisations want to protect their brand, ensure supply chain resiliency and choose trusted partners.

Solar said this new standard will create major supply chain disruptions.

“Because the same way that cybersecurity standards suddenly created a baseline of expectations so that organisations could demand certain standards from others. That's what's about to happen with AI,” she explained.

“Whether organisations are adhering to global standards and best practices is going to start to determine who you partner with and who you choose in your supply chain.”

She added, “This is where we're really trying to help Australia's commercial sector get ahead. Yes, there is a competitive advantage in innovation, but also there is a competitive advantage in adopting AI well.”

It is a risk not having AI

Embedding artificial intelligence in organisations may come with risks but the biggest danger is not implementing it at all, Solar told Digital Nation.

She said executives and boards need to embrace AI and realise there is an opportunity to unlock for their organisation if they navigate it responsibly.

“The things to consider, firstly, does an organisation know how they're using AI today. There's a lot of hidden use of AI as we're seeing with generative AI tools,” she said.

“They are so easy to use that employees are using them in the workplace anywhere between 30 to 40 percent, depending on which report you read, but most of that is hidden, so 68 percent of that generative AI usage is hidden.”

Once an organisation understands how it uses AI, leaders need to figure out their business strategy. Following that, comes governance, Solar explained.

“Implementing governance processes to start getting ahead proactively in shaping AI to be used responsibly across the organisation,” she said.

Solar noted one of the most important things for businesses to do is implement a generative AI policy, “It doesn't matter what they're standing on, the fact is employees are using it.”