What are the threats and opportunities of data science and machine learning?

  • Describe some of the challenges involved with the data used in machine learning
  • Identify how financial advisers' clients could be vulnerable
  • Explain how issues such as bias are being dealt with
What are the threats and opportunities of data science and machine learning?
Photo by Hitesh Choudhary from Pexels

As with most things, many of the threats and challenges in using data science, artificial intelligence and machine learning techniques can also be presented as opportunities, if not for today, then certainly for the future.

The starting point with much of this area of work is with the data itself. In order to train machine learning capabilities you need to have access to data – lots of it, and of sufficiently high quality in order to achieve the best results. We are all familiar with the phrase ‘garbage in, garbage out’.

We are incredibly privileged, working in financial services, to have access to reams of data far beyond that collected by many other sectors. The sensitive nature and depth of information held on clients and customers is immense.

Article continues after advert

Regulatory requirements – including Know Your Customer (KYC) – dictate that financial advisers gather a great deal of information in order to do their jobs compliantly and effectively. Some of this data is quite sophisticated, offering insights into clients’ wants, needs, goals and aspirations – far beyond the actual ‘financial’ information itself.

Client information of this nature is often difficult to acquire and can prove incredibly powerful. I am quite sure other industries would try to move heaven and earth to access even a sliver of such insights into their own customers. 

While such privileged access is one of our richest opportunities, it brings some potential concerns. For example, having sufficient data to feed into an AI model naturally favours larger players. The larger the player, the more data they can gather and the more honed their services can become.

As the technology underpinning many of these models allows such a business to scale even more dramatically, the advantage gap widens at a disproportionate pace and it becomes significantly harder for smaller players to compete with their larger peers. It is a common pattern in technology; the successful players scale massively, that scale brings advantages, those advantages allow even more scale.

One way to address this challenge is to place the ownership and control of personal data firmly in the hands of the individual. To that end, the introduction of open data and open finance standards such as Open Banking will be absolutely critical.

If we do not reach a scenario where we have open data, services and APIs (application programming interfaces), data – and its control – will be pushed outside the realm of the individual towards the hands of the large organisation.

Individuals own and control their own data and so should be free to share and revoke it with whomever they want. Having said that, we must ensure that individuals can share data in a secure, simple and transparent manner.

Advances in cloud-based technology, in combination with regulation and security, are evolving in such a way to tackle concerns. Ultimately, I see the collection and aggregation of data in our industry as an opportunity and a potential threat, the classic double-edged sword. The services we build using data can be of huge benefit to both the individual and society, as long as we continue to respect an individual’s data rights and build open platforms.