Boards urged to upskill and take AI governance seriously

Governance frameworks that have been applied to use of technology are unlikely to cut it for New Zealand organisations moving beyond experimenting with AI to deploying it.

That’s the message from the Institute of Directors which launched A Director’s Guide to AI Board Governance in Auckland today at its Governing AI conference. The guide outlines nine core principles to help boards effectively navigate the AI landscape, which the IoD says by nature requires a slightly different approach to deploying other technologies, such as cloud computing platforms or Internet of Things networks.

“AI learns and adapts over time, needing constant oversight and adjustments. Changes may disrupt consumer behaviour and stakeholder expectations, so organisations will need to stay flexible and well-informed,” the IoD’s guide points out. 

The IoD’s nine principles for governing AI

“AI can reinforce biases, lack transparency and make errors (hallucinations). Therefore, boards need to prioritise ethical considerations and accountability.”

This proactive approach is crucial, as the 2023 IoD Director Sentiment survey revealed a concerning gap: while nearly 70% of directors acknowledge technology will change boardroom operations, only 25% are focusing on “digital acceleration”, which AI is increasingly spearheading since the debut of generative AI services like ChatGPT.

The guide emphasises that AI is not just a fleeting trend but a powerful tool that can revolutionise operations, boost productivity, and ignite innovation. However, it also introduces complex risks that boards must carefully manage.

"The immediate challenge is to figure out how to harness this powerful tool in a way that is both productive and safe, and truly value adding," says Kirsten Patterson, Chief Executive of the NZ Institute of Directors.

Building board capability in AI is another critical theme. The guide stresses that while directors don't need to be AI experts, they should have sufficient knowledge to make informed decisions and engage in substantive dialogues. Continuous learning and staying informed about AI trends are crucial for effective governance.

Think of the workers

The IoD also addresses the impact of AI on workforce dynamics. Boards are urged to prioritise workforce wellbeing and continuous learning as AI becomes increasingly integral to operations. Clear communication about AI's role in the organization and providing opportunities for skill development are essential for maintaining a healthy organisational culture amid AI-driven change.

Speaking at the Governing AI conference in Auckland, Simply Privacy’s Frith Tweedie said that responsible AI gives organisations licence to continue innovating, while the flipside can be a loss of trust and social licence and unwelcome headlines.

She pointed to the case of a Māori woman who was mistakenly identified as a thief while shopping at a New World supermarket in Rotorua. Facial recognition technology, which Foodstuffs North Island is trialling, was responsible for the error, which resulted in the woman being approached by two staff members and being asked to leave. 

Tweedie said the unfortunate error was “no great surprise”. 

“I mean, they're improving all the time in terms of their ability to make matches, but most of them are coming from North American and Chinese vendors,” she pointed out. 

“Typically, they've been trained on North American data sets or Chinese data sets, and they're not necessarily well set up to cope with demographics typically found in New Zealand, which can lead to false positives, like we saw in that incident."

The IoD’s guide encourages boards to embrace AI as part of their governance practice. AI-powered tools can enhance board efficiency, from improving document sharing to providing data-driven insights for decision-making. However, the guide cautions that these tools should complement human judgment, not replace it.

Previous
Previous

Crowdstrike outage: Community thoughts?

Next
Next

Social robot or digital avatar, users interact with this AI technology as if it’s real