The policy for responsible use of AI in government requires departments and agencies to make a publicly available statement outlining their approach to AI adoption. And, as Jake Majerovic writes, the deadline is fast-approaching.
While you’re likely well-acquainted with the standard for AI transparency statements and have either completed version 1.0 or are well underway, the 28 February deadline means it’s time to go beyond ticking compliance boxes to measuring the impact of transparency.
This is where the real opportunity lies: ensuring that what you have in place is not only compliant but fit for purpose and assures the effectiveness of your agency’s AI systems.
It’s not just about meeting requirements – it’s about building public trust and protecting the public against negative impacts by fostering clear and accountable transparency.
The most effective transparency statements address critical questions, such as:
- have you done QA on your usage patterns and domains?
- how are your AI systems monitored to ensure ongoing effectiveness?
- are all deployed AI systems catalogued, registered and auditable?
- how can you prove compliance to applicable legislation and regulation?
These considerations aren’t new, but the depth and rigor of your responses are where strategic oversight becomes essential. Identifying risks isn’t sufficient. You must also demonstrate proactive measures that protect the public and align with broader compliance frameworks.
Public protection controls in AI governance
The trust and safety of the public must be at the core of every AI system deployed by government agencies. It’s about showing that your processes actively mitigate risks and protect those affected by your systems. You need to consider:
- what public protection controls are in place?
- how do you know what the negative impacts are – or could be?
- is there a risk register in place, and how is it maintained?
- how are you proving compliance with each requirement in the policy for the responsible use of AI in government?
- what tools do you use to control your AI transparency statements version, and who is responsible for managing their accuracy and currency?
Embedding transparency in broader governance standards
A strong transparency statement isn’t just a compliance obligation, it’s an opportunity to demonstrate alignment with this agenda while showcasing your agency’s commitment and accountability to public trust and protection.
To achieve this, agencies must actively measure trust – not just the existence of trust, but how it’s being fostered and communicated. This requires tools and methodologies to assess trust metrics and ensure these are reported in meaningful ways.
Transparency is foundational to the Australian public service reform agenda, and demands transparency statements not only be accurate, but also demonstrate continuous improvement. This means asking yourself:
- what tools are in place to ensure accuracy and currency over time?
- how are you measuring public trust in your agency or department, given that transparency is central to building it?
- what methodologies are used to assess public trust metrics, and how are these metrics communicated effectively?
- as part of the APS reform agenda, how are you evaluating the efficacy of your transparency statement against its objectives?
- how does your transparency statement align with the OECD definition of AI?
It’s also about strategic alignment. Your transparency statement must fit seamlessly into your agency’s broader governance frameworks and RegTech stacks, including compliance with AI policies and the responsible use of AI in government. It’s not enough to meet minimum criteria; your approach should actively demonstrate value to stakeholders and align with government priorities.
Clarity is key
Compliance is essential, but true transparency requires clarity. It’s not enough to simply make words visible. You need to ensure that the message is clear and accessible to all stakeholders, using plain, straightforward language. This is where agencies often need the most help.
Statements go beyond just ticking boxes. They must:
- meet compliance requirements while aligning with best practices
- use plain, clear language consistent with the Australian Government style manual
- build trust through clarity and human-centric communication.
A health check on your language is key to transforming your transparency efforts from a show-and-tell exercise into something that truly resonates. That’s how you build trust and provide effective protection, moving from merely complying with regulations to fully embracing the spirit of transparency.
As the 28 February deadline looms, this is the perfect moment to step back and assess your efforts. Agencies must use this opportunity to examine their existing transparency statements to ensure they add real value – not just to the public –but also to the agency’s internal governance processes.
Jake Majerovic – managing director Thinkless
Leave a Reply