Could developments in generative AI herald the beginning of the era of bionic advice: a hybrid of human judgement backed by an all-knowing and intelligent assistant adviser? Ihab El-Saie explores the issues.

The revolution in generative AI has triggered a frenzy of speculation about the future of regulated financial advice. The accessibility of the GPT engine through the freely available ChatGPT has had the powerful effect of convincing any remaining sceptics that change is upon us.

Could this herald the beginning of the era of bionic advice: a hybrid of human judgement backed by an all-knowing and intelligent assistant adviser? This will take advice to new heights, with real-time intelligent, engaging and personalised reporting without adding to the cost.

Robo advice: The story so far

Robo advice has used very little AI to date. It has mainly consisted of automating onboarding journeys and offering digital simplified model portfolio journeys so that little human intervention is required to set up a client.

Whilst streamlining some back office processes, supporting compliance and paraplanning and automating data capture are all useful, they hardly scratch the surface when it comes to the potential of generative, large language model-based AI.

‘Serious AI' has so far only been used in niche areas like fraud prevention and text recognition. The next wave of AI was set to be focused on gaining insights from large quantities of data using machine learning techniques. This is a still very much a live and productive area of progress.

However, natural language recognition and generative AI have jumped to the top of the queue with the game-changing ability to understand normal language and reply in perfectly formed grammar. This has been seen for a very long time as the distinction between human and machine. This is not so surprising when you consider that the neural network concept that this is based on, is directly inspired by the biology of the brain.

The next great leap forward

The public release of ChatGPT is like the AI starter's gun. The race is on. All of human knowledge, let alone financial advice will become more accessible. There is a lot of hype right now, just as there was a lot of hype at the start of the internet revolution. The question is whether it's justified.

The integration of the GPT engine into Microsoft's cloud development platform allows software developers to tap into this magical ability enabling software to understand and communicate. Critically, it allows for all personal data and conversations to remain secure and private. Compliance can breathe easy again.

The combination of training an AI engine on massive amounts of sector-specific data and its ability to act as an interlocutor opens up compelling possibilities. For instance, the Khan Academy, an American educational non-profit was afforded early access to the GPT engine by Microsoft. They have just announced that alongside their video-based tutorials, they are providing each student with a virtual AI tutor trained in their subjects and able to engage in normal conversations to help the student understand the subject matter.

It is now becoming apparent that even creative tasks such as poetry, image creation and humour are no longer exclusively the preserve of humans. The assistance that AI can provide in the future needs to be re-imagined so that it can bring optimum benefits for consumers, advisers and society at large.

Regulating AI

A few years ago, in a Facebook experiment, two AI chatbots were asked to negotiate a trade with each other and before long they had invented their own language and no one observing could understand what they were communicating between them. (Understandably, Facebook chose not to pursue this experiment.)

There are potential dangers in the use of AI, most obviously if used in a military context. The challenge is associated with machines that are more ‘knowledgeable' than humans and can process and absorb new data faster than them. Machines also have infinite memories, are able to control hardware and software, and communicate in normal language.

Many tech luminaries including Elon Musk are seriously concerned and have publicised their fears. Legislation will inevitably be required to regulate the use of AI and to attempt to circumscribe its use within safe and ethical boundaries to help rather than harm people.

In the context of financial regulation, humans will have to sense-check the AI's output for accuracy and compliance. ChatGPT for example has been known to make stuff up. It will have to be taught not to do that especially in the context of regulated advice. A combination of building trust and legislative consequences will be required to reign in and modify the more errant aspects of AI.

Generative AI: A game changer for advice

Although AI can be seen as a threat, it is more likely to result in great benefits to both the consumer and the adviser community. Advisers are always seeking tools to help provide a better service to their clients. AI offers the ultimate tool to turbo-charge the advice industry.

Imagine having multiple trained virtual assistants which can professionally deal with an adviser's grunt work and answer routine questions that your clients may have. Providing your clients with whole of market information and options optimised to suit each individual could become an almost effortless exercise.

This would free the regulated adviser to perform higher-level work, including relationship-building and understanding their clients' individual contexts. There are also implications here for meeting and exceeding new Consumer Duty obligations. AI-powered digital assistants take enough of the administrative work out of the hands of IFAs to enable them to increase engagement with their customers, aided and abetted by highly personalised illustrations to help paint a picture of what clients' savings might buy them in terms of lifestyle and opportunities to support the next generation.

At the very least, it makes it possible for many more consumers to access some level of well-informed financial advice, enabling them to explore and understand the options available to them to optimise their retirement income right through their lives and even beyond.

The adviser will have the time to build a rapport with their client and will lean more heavily than ever on informed intuition, judgement, empathy and face-to-face interaction to supplement the plethora of accurate information and engaging reports and illustrations that the AI will be able to generate.

AI finally forces us to consider what is uniquely human. Work is about to become much more interesting. Routine tasks will be delegated to digital assistants who will never complain, take sick days or holidays. Sure, these assistants may require some upgrading and patching every now and then to keep them secure and functioning properly. However, the workload they will take on will only grow once we have worked out what activities to delegate to them usefully.

Considering all of the above, it doesn't take a brave person to predict that AI will be able to help regulated advisers to offer a much higher quality and broader service and will finally help us address the advice gap which has been steadily growing over the last ten years since the Retail Distribution Review went live. The age of the bionic adviser is approaching with its human masters and armies of AI workers.

Ihab El-Saie is chief executive at Dunstan Thomas