This product has already been purchased.
- AI fluency for directors is not about technical coding skills. It’s the ability to interpret data dashboards, ask sharp questions and effectively challenge management claims.
- Governance requires leveraging existing director capabilities – strategy, risk and financial accountability – to assess AI use cases, “right-size” risks and verify outcomes. Avoid relying on a single “AI director”.
- Fluency must be an ongoing, whole-of-board habit to ensure safe, value-driven technology adoption.
AI governance doesn’t require becoming an expert, it demands the same rigorous scrutiny that is applied to strategy and finance. Here is how directors can build essential, actionable AI fluency.
If a boardroom discussion on artificial intelligence (AI) seems overwhelming, don’t panic. There are ways to boost your AI fluency so that you can effectively contribute to sound AI decisions.
Katherine Boiciuc GAICD, EY Oceania Chief Technology and Innovation Officer, says AI fluency isn’t technical literacy. “Technical literacy is understanding how an AI model is built. Fluency is knowing what ‘good’ looks like and being able to tell when you’re not getting it. It’s the difference between building a car and driving it. A director’s job is to read the dashboard and know when to pull over, not to build the engine.”
Boiciuc says you don’t need to write code to govern AI any more than you write the ledger to govern an organisation’s finances. “You need enough to ask a sharper question and recognise a weak answer.”
Professor Nicholas Davis, Industry Professor, Emerging Technology and Co-director at the University of Sydney Human Technology Institute (HTI), says AI isn’t an area where directors suddenly need an entirely new technical skill set. Instead, he suggests you should transfer the skills needed to be a competent director to AI decision making.
For emerging directors, he says, many of these competencies are gained by completing the AICD Company Directors Course, as well as understanding your obligations under the Corporations Act and what a well-performing board looks like.
If you aren’t digitally savvy, he advises building a “minimum viable understanding” of AI to fulfil your duties, without trying to become an expert.
Like Boiciuc, he stresses the importance of questions. “Don’t be afraid to say, I don’t understand. If you still don’t understand the answer given, ask for it to be explained again in a different way.”
Davis adds, “AI fluency is about understanding the consequences of an AI decision, both positive and negative, and being able to link that back to your director’s duties and areas like strategy, risk management and financial accountability.”
Define the problem before getting the tool
Not having the time or resources to master AI governance should not hold back emerging directors, even those on small or not-for-profit (NFP) boards.
The first step, according to Boiciuc, is to take the pressure off yourself. “You’re not becoming a data scientist, and you don’t need enterprise-grade governance to govern AI well. Right-size it to your risk.
“Start with the problem, not the technology. Where could AI help your mission, and where could it quietly harm those you serve? Use the tools yourself, because an hour a week teaches you more than any briefing. Ask management for one page on where you use AI, on what data, with what risk and who owns it. Every time an AI issue comes up, add it to a ‘questions to ask management’ checklist .”
For Davis, a good start is the organisation’s biggest pain points. “You don’t need to know everything about AI, you just need to be able to interrogate the potential set of solutions to solve a pain point.
“Ask for outside help where it’s needed, for example, from directors in the community, organisations like HTI and others in the university or NFP space. Places such as Infoxchange support NFPs, while the National AI Centre helps small and medium-sized enterprises.”
Davis says AI can be daunting for many directors because it’s changing so rapidly. “Boards should have a standing item every couple of meetings just to check in on where they are on AI and what they’re worried about.”
Take small steps
Lee Hickin, executive director of the National AI Centre, advises directors with limited time and resources to move in small steps, in tandem with risk management and opportunity exploration.
He says they should look at what others in their industry are doing in Australia and similar countries. “You can also borrow from lots of different sectors. Attend industry presentations and those for directors, then watch more widely what’s happening. To broaden your horizons, you should receive information from multiple places.
“Look for things your competitors may not see and spend as much time considering the opportunities as you would the risks.”
Keep asking questions
“A board that can’t question a management claim ends up accepting it. That’s where the risks start,” says Boiciuc.
“Ordinary automation gets badged as AI and best-case business cases go unchallenged, so you sign off on value that never arrives. In addition, board papers are increasingly AI-drafted, so real decisions can be made on facts that haven’t been properly verified.”
Davis warns that not asking the right questions can lead to a misalignment between board and management thinking on AI. It can also enable vendors or management to push forward with expensive AI projects that don’t return the value promised.
Using vs governing
Personal AI use will not automatically equip you for boardroom AI decision making. Boiciuc says using AI is a skill, governing AI is a duty, and directors need both.
“Personal use builds calibration. You learn where these tools work well and where they quietly fail, and it makes you harder to bluff. But using it daily isn’t the same as understanding the risk when a model makes thousands of decisions on your data, at scale,” he says.
Davis agrees, noting that personal use of AI tools is often not a good indicator of the potential for AI at an organisational level.
Using AI for personal administration and communication can save someone time. However, Davis says the data shows that doesn’t translate into organisational productivity, which is about coordinating many people, workflows, and processes.
“An individual director using AI may not be able to effectively think about all the different ways AI could go wrong when 1000 people use it, or how outside contractors or third parties may use it differently,” he says.
The bottom line
Boiciuc stresses the significance of gaining confidence in using and talking about AI. “It’s the thing that makes directors defer instead of challenge. Improving that will lead to sharper questions, better scrutiny and faster, safer adoption.
“Remember that fluency is perishable, so treat it as a habit, not a course you complete. Don’t outsource it to one person. The ‘AI director’ is a trap. The uplift has to be whole-of-board. AI doesn’t dilute your duties. It raises the bar.
“Govern the data and decisions, and most of the AI risk becomes manageable. Keep both risks in view. The downside is obvious, but so is the cost of doing nothing.”
Latest news
Already a member?
Login to view this content