AUSTRALIA Law and Practice Contributed by: Andrew Stone, Dhanushka Jayawardena, Andrew Choi and Chris Kinsella, Holding Redlich
of AI within investment management. Relevant laws and harmful use cases include the following. • AI system is not sufficiently secure – laws relating to directors’ duties are relevant, including the duty to exercise powers with due care and diligence in order to assess and govern risks to the organisa - tion (including non-financial risk). • Misleading outputs or statements – Australian consumer and other laws that prohibit engaging misleading and deceptive conduct and making false and misleading representations may apply, for example, if: (a) outputs are provided by AI and then included in a disclosure document issued by a fund sponsor; or (b) a fund sponsor makes: (i) misleading representations or remains silent as to when AI is being used; or (ii) misleading statements as to the perfor - mance and outputs of the AI systems. • Harmful outputs – trust laws applying to fiduciaries and the laws of negligence may apply in relation to harmful outputs produced by AI if the person using the output owed relevant duties to the person who suffers loss or harm as a result of the output. • Misuse of data or infringement of model or system – privacy laws, IP laws (including copyright), duties of confidence and contract protect the use, repro - duction and/or disclosure of data (including training data, input data, and outputs) and the model or system without the requisite consents or rights. Privacy laws restrict the collection of personal information for an improper purpose and impose transparency and data minimisation requirements on the handling of personal information. • Bias, incorrect or poor-quality output – privacy laws impose quality and accuracy obligations that may apply to training and input data (that is personal information) and outputs (where new per - sonal information is generated). Given the above, AFSL holders must: • do all things necessary to ensure that the financial services covered by the licence are provided effi - ciently, honestly and fairly;
• have adequate technological resources to provide the financial services covered by the licence and to carry out supervisory arrangements (for example, of representatives of the licensee); and • have adequate risk management systems. Each of these obligations is relevant to the manner in which an AFSL holder may choose to engage with AI. Within this context, the Australian government issued the Voluntary AI Safety Standard (VAISS) in September 2024. The VAISS includes ten voluntary AI guardrails to help organisations deploy and use AI systems with - in the bounds of existing Australian laws, emerging regulatory guidance, and community expectations. The guardrails are as follows: • establish, implement and publish an accountability process including governance, internal capability, and a strategy for regulatory compliance; • establish and implement a risk management pro - cess to identify and mitigate risks; • protect AI systems and implement data govern - ance measures to manage data quality and prov - enance; • test AI models and systems to evaluate model per - formance and monitor the system once deployed; • enable human control or intervention in an AI sys - tem to achieve meaningful human oversight; • inform end users regarding AI-enabled decisions, interactions with AI, and AI-generated content; • establish processes for people impacted by AI systems to challenge use or outcomes; • be transparent with other organisations across the AI supply chain about data, models and systems to help them effectively address risks; • keep and maintain records to allow third parties to assess compliance with guardrails; and • engage stakeholders and evaluate their needs and circumstances, with a focus on safety, diversity, inclusion and fairness. Adopting these guardrails will create a foundation for the safe and responsible use of AI. It will make it eas - ier for any organisation to comply with any potential future regulatory requirements in Australia and emerg -
19 CHAMBERS.COM
Powered by FlippingBook