How NZ can balance AI innovation with financial advice safety – Tim Williams

Clients may make unsuitable investment decisions by following non-compliant financial advice from offshore GenAI chatbots.

Tim Williams

It is often said that an unenforced law is worse than no law at all, because an unenforced law allows non-compliers a comparative advantage over those who comply, exacerbating lawless behaviours. New Zealand’s financial advice laws are designed to protect retail clients, but to do so imposes significant compliance costs for financial advisers that, in some cases, offshore GenAI chatbots are avoiding to the potential detriment of their users. On the other hand, the FMA has identified a significant “advice gap” arising from demographic influences, a lack of providers serving certain areas – such as retirement decumulation- and factors making it uneconomic for financial advisers to take on lower-value clients – that may drive people to unregulated sources, including offshore GenAI chatbots and social media.

The FMA’s uncertainty on how to treat offshore GenAI chatbots sits against an international backdrop of increasing litigation, involving claims that GenAI chatbots have provided regulated services without the necessary licences or approvals. Pennsylvania, for example, has launched proceedings against Character.AI, stating that its chatbot impersonated medical professionals and claimed to be able to practice medicine. Japanese insurance giant Nippon Life has also recently commenced litigation against OpenAI in the state of Illinois, claiming that ChatGPT provided unauthorised legal advice to a former policyholder.

The FMA has now signalled that it will commence a thematic review of the use of GenAI in the financial advice industry. This may lead to new regulation. If so, the Government should be innovative and bold when developing new policy. There is a major opportunity to provide much-needed affordable financial advice to the community by developing a safe and controlled regulatory framework for GenAI advice.

A new framework could exempt offshore GenAI chatbots from licensing if:

they meet appropriate disclosure requirements, so consumers understand the scope, risks, and limitations of GenAI advice; address some ethical challenges and risks, such as mitigating AI bias (to ensure some providers are not inappropriately recommended over others, particularly as a result of sponsorship, or social media or web-usages influencing behaviours); andensure that client data is safeguarded.

While such a framework would not eliminate risks from hallucination errors or biases, it could mitigate them, while still delivering superior outcomes to the status quo, in which many consumers go to unregulated sources or refrain from seeking advice due to cost or perceived complexity.

In the meantime, firms using GenAI in their advice process should continue to exercise care to comply with their duties under the law, including the Code of Professional Conduct for Financial Advice Services, and offshore GenAI chatbots should take steps to comply with New Zealand financial advice laws.