ASIC has conducted a report into how AFSLs are starting to use artificial intelligence (AI) in their practices and possible areas where they could come unstuck from a governance perspective.
In the regulator’s report Beware the gap: Governance arrangements in the face of AI innovation, it spoke with 23 licensees on 624 cases of AI usage. Licensees, both AFSLs and credit licensees, were asked about their risk management and governance arrangements as well as how they planned to use AI in the future.
Overall, ASIC said licensees’ use of AI is “quite cautious” currently but is increasing rapidly, as 61 per cent of licensees said they want to increase their usage of AI in the next 12 months.
Usage is currently centred around well-known and established techniques, but it has observed an increase in complex techniques such as deep learning and generative AI (GenAI) which represent around a third of cases. In particular, GenAI is expected to increase “exponentially”, ASIC said.
The most common uses, as of December 2023, are:
| Area | Usage |
| Marketing communication | Analysing consumers’ spending patterns for marketing purposes. Optimising marketing communication. |
| Customer engagement and customer value proposition | Chatbots to answer simple questions. Cash flow forecasting and budgeting tools. |
| Fraud detection | Transaction monitoring and identification of fraudulent documents. |
| Business efficiencies and compliance | Call transcription analytics. Triaging incoming complaints. Document indexing. |
| Pricing optimisation |
Assisting in determining discretionary offers. Targeted retention offers to reduce client switching. |
| Insurance | Supporting the claims process. Actuarial models for risk, cost and demand modelling. |
Risk management
ASIC said it observed gaps in arrangements for managing AI risk and some licensees opted to address AI risk from the perspective of the business rather than the consumer.
Approximately half of licensees had specifically updated their risk management policies or procedures to address AI risks.
Nearly half of licensees did not have policies in place that considered consumer fairness or bias, and even fewer had policies governing the disclosure of AI use to consumers.
It also found the maturity of risk and governance did not always align with the nature and scale of licensee’s AI usage which created the greatest risk of consumer harm. A future challenge will be for licensees to maintain the adequacy of their arrangements and ensure they are fully operationalised as their AI use grows in scale and complexity, particularly if their approach to AI governance is already fragmented.
Joe Longo, ASIC chair, said: “We want to see licensees harness the potential for AI in a safe and responsible manner – one that benefits consumers and financial markets. This can only happen if adequate governance arrangements are in place before AI is deployed.
“Without appropriate governance, we risk seeing misinformation, unintended discrimination or bias, manipulation of consumer sentiment and data security and privacy failures, all of which has the potential to cause consumer harm and damage to market confidence.”
Next steps
ASIC outlined multiple steps for AFSLs to consider when it comes to enacting their AI usage:
- Licensees must do all things necessary to ensure that financial services or credit services are provided in a way that meets all of the elements of “efficiently, honestly and fairly”.
- Licensees must not engage in unconscionable conduct.
- Licensees must not make false or misleading representations.
- Licensees should have measures for complying with their obligations, including their general obligations.
Licensees must have adequate technological and human resources. - Licensees must have adequate risk management systems.
- Licensees remain responsible for outsourced functions.
- Company directors and officers must discharge their duties with a reasonable degree of care and diligence.
