Work Better Together

My Account

Want to hear my thoughts on these topics and more? Subscribe to my newsletter here

Institute of Directors conference 2024 Top 5: AI for directors

My musings on the comments I heard on AI over the Institute of Directors conference in May 2024 (check out my other top 5's too). The panel discussion (Matt Ensor, Liz MacPherson, Alistair Miller, Hillary Walton) on AI at the IoD conference was fascinating to a low-tech person like myself. I found myself particularly stuck on the idea that we need to learn how to become competent in using the appropriate AI to do tasks, not how to do the task itself. But how will we monitor in the future whether the task has been done correctly or what impact bias and assumptions have on the finished product? This a different conversation to when we moved to using a calculator to do sums – how do we build in ethics and nuance where there is more than one right answer? I found the following points interesting to consider:

- What is the AI strategy for your organisation or how can you source the expertise to create one?
- Your people are already using AI, they may just not realise it (Grammarly, co-pilot etc). You need a policy to guide usage and education to ensure employees are making good decisions about how to fit use with strategy (is it a distraction or moving you towards your targets?). Is it possible to have an organisational register listing everywhere AI is being used – hidden and overt?
- AI tools should not be given a whole view of the organisation, segment it for specific purposes and only use a private tool for work related information.
- Overlay AI on top of the current risk profile – what falls outside the current controls? Privacy impacts need to be assessed on the AI tools that are used, not just the inputs to the tools.
- How does this align with your organisation’s approach to Māori data sovereignty – call this out in your strategy.

Extra resources I'm going to read!
https://privacy.org.nz/news/media-releases/privacy-commissioner-outlines-expectations-around-ai-use/