NIBA / Special Feature
AUSTRALIA’S AI ETHICS PRINCIPLES
“Consumers will ask, ‘did I consent to my data being used in that way’ – and that notion of consent is huge when it comes to the ethical use of AI,” says Dr Neary. Know What You’re Getting In To When assessing AI platforms, and then the ethical use of those platforms, there’s plenty to consider – including the cybersecurity credentials and governance structures of providers. Weaver says, “Vetting should be risk- adjusted – the more sensitive the data or critical the outcome – the more rigorously businesses should review. With high sensitivity data – where data is hosted – the cybersecurity standards of all providers and governance are vital. “For critical use cases, businesses should clearly understand the function of the AI, whether it's been tested for similar use cases before and ensure it's trained with a relevant training set. Whatever upfront due diligence is done – it’s just as important to ensure performance is tracked and reviewed to ensure it performs as expected.” The quality and correctness of data being used in an AI platform is also fundamental to get right. “If you’re going to use AI it is imperative that you have the right data strategy and controls in place. If you start off with even slightly faulty data, it’s going to impact the integrity of what you get out” says Dr Neary. Time to Take a Beat – and Think! Using AI in your own business is one thing – your clients using AI, and having to present that risk to insurers, is another. And that in itself is an ever evolving area of consideration. Weaver says, “AI-related exposures often fall between cyber, general liability, and tech E&O policies. Third-party claims may be covered, but it’s unlikely an SMEs own losses from AI failures will be.
The Department of Industry, Science and Resources has published Australia’s 8 AI Ethics Principles, which are designed to help businesses implement AI safely, securely and reliably. They are:
Human, societal and environmental wellbeing AI systems should benefit individuals, society and the environment. 1
Human-centred values AI systems should respect human rights, diversity, and the autonomy of individuals. 2
4
3
Privacy protection and security AI systems should respect and uphold privacy rights and data protection, and ensure the security of data.
Fairness AI systems should be inclusive and accessible, and should not involve or result in unfair discrimination against individuals, communities or groups. Reliability and safety AI systems should reliably operate in accordance with their intended purpose. 5 Contestability When an AI system significantly impacts a person, community, group or environment, there should be a timely process to allow people to challenge the use or outcomes of the AI system. 7
6
Transparency and explainability
There should be transparency and responsible disclosure so people can understand when they are being significantly impacted by AI, and can find out when an AI system is engaging with them.
8
Accountability People responsible for the
different phases of the AI system lifecycle should be identifiable and accountable for the outcomes of the AI systems, and human oversight of AI systems should be enabled.
48 / INSURANCE ADVISER OCTOBER 2025
Made with FlippingBook Digital Publishing Software