The rapid growth in the popularity of digital or “robo” advice has led ASIC to release new guidance on the regulation of digital advice in draft regulatory guide Providing digital financial product advice to retail clients.
The guide has implications both for providers of digital advice and for all Australian financial services licensees.
While we welcome this guidance on the issues currently facing providers of digital advice, we encourage ASIC to think ahead to the potential impact of future technological developments in this space. The integration of cognitive technologies and gamification tools with digital advice raises interesting questions about how the best interest duty can be met, whether the traditional distinction between general and personal advice remains relevant, and whether there are more effective methods of disclosure through digital channels.
Key takeaways from the new regulatory guide for digital advisers
The draft regulatory guide sets out ASIC’s views on a number of regulatory issues digital advice licensees face. Some key points are identified below.
- Maintaining organisational competence – digital advice licensees need to have at least one responsible manager who meets the minimum training and competence standards for advisers. ASIC has granted existing licensees a 6 month transition period to comply with this requirement.
- Having adequate human resources to provide financial services – digital advice licensees need to:
- have at least one person with an understanding of the technology and algorithms used to provide digital advice (this does not require all licensees to understand the specific computer coding of an algorithm, however ASIC expects that there are people within the business who understand the rationale, risks and rules behind the algorithms underpinning the advice); and
- have at least one person with the appropriate skills and experience to thoroughly conduct regular reviews of the quality of digital advice to ensure it is legally compliant.
- Monitoring and testing algorithms – digital advice licensees must monitor and test their algorithms through a documented test strategy, being able to control, monitor and reconstruct any changes to algorithms over a seven-year timeframe.
- Scaled advice – customers must demonstrate they understand the advice they want is within the scope being offered by a digital licensee. ASIC recommends a “triage” process is adopted, filtering out those clients for whom digital advice is not suitable and who do not demonstrate they understand the scope of the advice being provided. Importantly, ASIC notes that a digital advice model that results in all clients receiving advice would raise serious concerns and would prompt close scrutiny from ASIC.
- Compensation arrangements – digital advice licensees should consider the potential for widespread loss if an algorithm is flawed and the aggregation of claims clauses in professional indemnity insurance policies.
Cyber security – an issue for all financial services licensees
ASIC views adequate data security as part of demonstrating that a licensee has adequate technological resources. ASIC expects all licensees to have in place security compliance measures, such as Cloud computing security for tenants published by the Australian Cyber Security Centre, or equivalent. For digital advice licensees, ASIC expects that cyber security and information security protocols are assessed against recognised frameworks.
Looking forward to future developments in digital advice
Digital advice will continue to evolve as companies look increasingly to leverage the benefits of data and cognitive technologies expand in scope and application. The draft regulatory guidance should consider the impact that these developments may have on the digital advice space in the future.
How is information “collected” from a potential client?
The traditional advice model requires a human adviser to collect information from a client, generally through a questionnaire or an adviser asking a series of questions. This traditional model is recognised in the current drafting of the best interest duty safe harbour, which requires an adviser to identify the objectives, financial situation and needs of a client that were disclosed to the adviser by the client through instructions to discharge the obligation to act in a client’s best interest.
However, collection of information directly from a client may not be the best or only way an adviser can obtain information about a client. Creators of gamification tools suggest that short investment-style games are more accurate in determining an individual’s risk profile than that individual’s perception of their own risk. Cognitive technologies already used overseas in a number of different contexts (including to assess a person’s credit worthiness) may also be used in the future to compile information from a person’s online presence and other unstructured data sources to build a profile of their objectives, financial situation and needs.
When digital advice is combined with cognitive technologies and gamification tools, there will be limited need for the client to give instructions about their objectives, financial situation and needs. The digital adviser will already be able to determine this based on other data sources, generating efficient, timely advice that has the potential to consider a range and depth of data that cannot be processed by a human adviser.
Given these anticipated developments in technology, it would be helpful for ASIC to comment on whether the best interest duty safe harbour can be met by digital advisers taking advantage of the benefits of cognitive technology and gamification tools.
Using data analytics to build a hypothetical client
The distinction between fact, general and personal advice will continue to be difficult to determine in the digital environment.
With developments in data analytics, we can see a future where digital advisers can create a hypothetical client who is of a certain age, gender, profession, marital status and residential area. Using this information, a digital adviser could ascertain the typical financial situation, needs, objectives and risk appetite of a client, based on data from previous clients who share similar characteristics.
Advice given in this context may not be personal advice, as advice has been given based on a theoretical client rather than a real client’s actual situation, needs and objectives. This raises interesting policy questions - should we continue to maintain a general/personal advice distinction or should we move to a sale/advice distinction, and what should be the trigger for the best interest duty? Although ASIC has provided some examples of the distinction between general and personal advice in the draft regulatory guide, it would be helpful if more examples were provided on this distinction in the digital advice context.
Effective and efficient disclosure in the digital environment
The current disclosure regime works best in a traditional face to face environment. In an online environment, clients will not spend the time to read lengthy disclosure documents. Given the rise in digital distribution channels it may be time to rethink how information is provided to clients.
In an online environment, clients could engage in an interactive form of progressive disclosure through the use of artificial intelligence platforms (similar to Amelia, Watson and Connie), or videos and pop up boxes that provide more information on complex issues. This type of progressive disclosure may promote better understanding as information will be given in context rather than in a standalone document. We welcome ASIC’s views on the potential interplay between artificial intelligence technology and disclosure, and how disclosure requirements can be met in an interactive way.