|
The Three Laws of Robo-advice * 1. The same laws apply to robo advice as human advice. |
On 21 March 2016, ASIC released the following:
· Media release 16-082MR.
· Consultation Paper 254 Regulating digital financial product advice (CP 254); and
· Draft Regulatory Guide on Providing digital financial product advice to retail clients.
In simple terms, ASIC’s proposed guidance can be summarised in the above three ‘laws’ (* with apologies to Isaac Asimov).
The key message is that ASIC confirms that the regulation of financial product advice is technology neutral and does not prevent the provision of digital advice generated by automated systems and algorithms. In particularly, there is no need for an individual adviser to review or take responsibility for each piece of advice generated by the system for clients.
ASIC is seeking comments from advice licensees and representatives, and other parties who are interested in the provision of digital financial product advice. The closing date for submissions is 16 May 2016. The final regulatory guide is due for release in August 2016.
CP 254 sets out ASIC’s proposed approach to the regulation of digital financial advice for retail clients:
· Digital advice licensees must have at least one responsible manager (RM) who meets the minimum training and competence standards for retail advice (currently set out in (RG 146). They must be responsible for the provision of digital advice by the licensee.
o This is required to meet the organisational competence and training requirements in section 912A(1)(e) of the CA.
o Digital advice is generated by algorithms and there is no need to have an individual review of each piece of digital advice given.
o ASIC has sought feedback whether the RM should meet the proposed new higher training and competence standards (i.e. have a degree or equivalent, pass an exam, complete a professional year, undertake continuing professional development and comply with the proposed ethical standards).
o ASIC has proposed a six month transition period to comply with the RM requirement to assist AFS licensees wanting to operate a digital advice business.
· Digital advice licensees must monitor and test the algorithms that underpin the advice and propose to issue guidance on ways in which digital advice licensees should monitor and test algorithms. ASIC’s expectations are:
o Appropriate system design documentation that clearly sets out the purpose, scope and design of their algorithms. Decision trees or decision rules should form part of this documentation where relevant.
o Documented test strategy that explains the scope of the testing of algorithms. This should include test plans, test cases, test results, defect resolution (if relevant), and final test results. Robust testing of algorithms are to occur before advice is first provided to a client, and on a regular basis after that.
o Appropriate processes for managing any changes to an algorithm. This includes having security arrangements in place to monitor and prevent unauthorised access to the algorithm.
o Ability to control, monitor and reconstruct any changes to algorithms over a seven year time frame (Note: in relation to personal advice to retail clients, ASIC is currently consulting on proposed amendments to Class Order [CO 14/923] Record-keeping obligations for Australian financial service licensees when giving personal advice to ensure that records are retained for seven years).
o Review and update algorithms whenever there are factors that may affect their currency (e.g market changes and changes in the law).
o Controls and processes to suspend the provision of advice if an error in an algorithm is detected.
o Adequate resources, including human and technological resources, to monitor and supervise the performance of algorithms through an adequate and timely review of the advice provided to clients.
o Robust compliance arrangements in place to monitor and test the quality of advice which should be reviewed by a human adviser for compliance with the law, and should not be a ‘tick-a-box’ exercise. Frequent reviews should be conducted initially with heightened scrutiny when any change to an algorithm is made and should be regularly monitored and tested through periodic and random advice reviews.
