AI ‘inevitable’ in pension master trust era – MHC

Large institutional investors, including pension schemes, are beginning to use artificial intelligence (AI)-based platforms and machine learning models, a Mason Hayes and Curran webinar has heard (7 September).

These tools are used to assist in long-term portfolio management in place of more traditional models based on financial theories.

Accurate planning

Fund providers have also begun using machine learning to predict scenarios such as an early withdrawal of the fund’s resources, which will allow for accurate planning.

When it comes to individual retirement planning, AI-based platforms can collect and analyse data on debts, assets, income, existing pensions and years left to retirement to determine the best plan of action.

On risk management, AI may provide pension trustees with real-time analysis of the impact of recessions, inflation or employer insolvency.

The identification of patterns and the probability of certain events will assist trustees and investment managers when it comes to making investment decisions, the webinar heard.

Chatbots are becoming a common feature for many online retailers and pension member queries could be dealt with in a similar manner, with a resulting reduction in administrative costs.

The era of the master trust and upcoming automatic enrolment will see schemes with massive membership numbers across thousands of employers.

That level of member data and employer interaction mean AI is almost inevitable, the webinar heard.

However, scheme and member data will need to be accurate and consistent.

This is likely to be a major application for AI, where it can identify patterns of behaviour based on data and allow for informed trustee decision-making.

Limitations

Apprehension about AI remains, and the reality is that it is unlikely to replace human administrators or investment managers in the near future.

However, AI can assist in taking on boring or repetitive tasks where the data provided is good enough.

This will allow staff to focus on more complex tasks that will always require human reasoning and intuition,  the webinar heard.

AI is likely to face many hurdles particularly on ethical grounds, where its adoption leads to job losses, and significant levels of AI regulation are likely in the coming years.

In the EU, the AI Act, due to be signed later this year, will introduce substantial regulation of AI systems on a sliding scale of risk.

The authority and expertise of regulators to pronounce on artificial intelligence remains a big question, the webinar heard.

Financial services’ legislation will combine with AI tech and pensions’ legislation, making a complicated whole, MHC pensions partner Stephen Gillick said.

Human advisors

Human financial advisors and brokers will always be needed, however, though menial tasks will be better automated.

Richard Skinner (CTO OF Marshmallow) said that mobile, digital pensions will shortly evolve as neo-bank accounts have, with the invention of fintech.

Users will be able to ‘chat’ with their funds and be presented with relevant information, rather than waiting for an annual report on the state of play.

“Tailored recommendations and tailored risk assessments and automatic rebalancing are technologically possible now,” he said, though oversight is needed to ensure the correct information is given to scheme members.

Innovations

However, the threat of AI wiping out jobs is not borne out by technological innovations over the past 50 years, which have been huge job creators.

“People just move on to spending time on more important tasks,” said Gillick.

A participant survey found that 74% would not be happy taking financial advice from an AI robot, while 26% would.

However, a resounding 72% agreed the integration of AI systems into the pensions industry has the potential to deliver better outcomes for pension scheme members.

And 91% of those surveyed are currently using AI in their pensions’ business.