Potential for AI-driven attacks on financial adviser databases elevates risks posed by ChatGPT

The problem created by the AI boom is that fraudsters are quickly proficient in adopting the technology.

Australia’s wealthiest investors are so concerned about privacy security, in the wake of hack attacks on Perpetual and Medibank, that the breakthrough promised by artificial intelligence could be severely restricted.

Financial advice is one of the standout sectors set to benefit from lower costs offered by AI tools such as ChatGPT, but the heightened threat the same technology poses to family fortunes held in financial adviser databases means the much hyped technology breakthrough could be delayed dramatically.

“There has been an escalation in concerns among investors about data security. The big corporate breaches we have seen recently have alerted people to cyber risk. The most common question now from clients is why do you need this information?” said Glenn Calder of Viridian Financial, speaking at this week’s annual Top 100 Advisers Summit. (New York-based Barron’s in conjunction with The Australian published the Top 100 Advisers List in September).

The financial advice sector has been hoping that AI tools will cut costs and lift productivity – a move that could make advice more affordable to the wider community. Top advisers believe the savings from AI on back-office work could slice the time taken to administer routine advice and related costs.

Many of the industry’s top adviser groups already have specialist working groups trialling ChatGPT, Google’s Bard and Microsoft’s Azure tools for back office-duties. Industry analysts suggest the costs of core financial advice services, such as super contribution plans, could be cut to a quarter of what they are today with AI.

But the flip side of the AI boom is that fraudsters are just as quick – if not quicker – to adopt the new technology. AI-driven fraud activity threatens to be considerably more powerful than the scams we know today.

One certain repercussion of the elevated risk in IT security from having AI inside advice organisations – or any business – will be the disadvantage felt by smaller competitors.

Already the upper end of the Top Advisers list is dominated by a handful of institutions such as Morgan Stanley, Macquarie Private Bank, Pitcher Partners and Koda Capital.

Alternatively, high ranking firms are ultimately part of a larger group: Shadforth Financial Group is part of Insignia Financial while Escala Partners is part owned by New York-based Focus Financial Partners.

As Calder warned: “Everyone in advice that is not equipped for this new technology will be replaced, the best AI users will be the best competitors.

“In something like portfolio construction for investors we are finding the AI models are getting it substantially correct; I’d say they are at least 80 per cent accurate. But then we have the continual risk of cybersecurity.”

David Simon of Integral Private Wealth said: “In terms of technology this is equivalent to the arrival of the internet, but we have to proceed very carefully.

“The way we are doing that is by only using AI for generic work – we are not using the real names of any clients. In no way are we putting clients at risk.

“Clearly this issue is crucial to every financial advisory firm and we also need to be careful about employees who are working from home.”

The financial advice sector in Australia has not yet reported a substantial data breach. The two most high-profile data breaches have been at Medibank and Optus.

However, inside financial services, the incident that has caught attention occurred at Perpetual. In June this year the group revealed personal data belonging to thousands of customers had leaked through a third party service provider – clients of the group were cut out from accessing the myPerpetual platform at the time.

Perpetual corresponded with 40,000 clients over the incident.