Brokerage head questions big data, AI and robo-advice

He said: “In recent months, hordes of financial analysts and market commentators have jumped on the bandwagon to exclaim how bright our futures will all become when we willingly submit to a life of automation.

“And who’s to say they’re wrong, as big data scientists and technologists have convinced even the biggest companies across the globe to invest millions of dollars into this at the expense of upskilling their own people.

“My question is: Why are we submitting to what we are simply being told and don’t really understand?”

Mr Russell said that while big data has the capacity for predictive decision making, it is only relevant in situations where “nothing changes”.

Highlighting that “major banks are diverting hundreds of millions of dollars away from their front-line customer service into big data”, he argued that “what we are not being told is that today more than ever before in our history, past trends and behaviours are actually proving to be quite unreliable in predicative modelling”.

The MoneyQuest MD elaborated: “The simple reason for this is that technology is evolving at such a rapid pace and delivering so many unknowns that our consumption behaviours are themselves becoming unreliable. This extends to our workplaces, our employment types, our housing situations, our relationships, how we communicate and so on.

“Big data as we know consists only of quantitative inputs and not the crucial qualitative inputs that make up our individual DNA ... our propensity to change and our individual stories — past, present and future.”

Further, Mr Russell said that algorithms can be used to make “really fast decision”, but said that they are not always the right ones.

He stated: “The one thing we know for sure about an algorithm is that we know nothing about that algorithm ... we are always prohibited from checking the truths in today’s algorithms.

“This will never change as all algorithms are based on the personal subjectivity and bias of their authors.”

Mr Russell suggested that credit-scoring algorithms used by banks and non-bank lenders to determine someone’s creditworthiness are “subjectively engineered using only quantitative [data]”.

He continued: “[I]ncreasingly, we are seeing many creditworthy applicants fail lenders’ auto credit-scoring, only to be salvaged once their story can be shared human to human, for in these stories we, as mortgage professionals, can properly assess an individual’s true character, capacity and credit risk.

“In the case of mortgage applicants, we know that someone’s past is not always an accurate indicator of their present or future, and if left unchecked, auto credit-scoring has the potential to be unwittingly unfair, unethical and discriminatory.”

‘We need to be smart and inquisitive’

As such, Mr Russell suggested that it is “imperative we need not only be a smart generation but, more than ever before, an inquisitive generation”.

Lastly, the MoneyQuest managing director said that the “jury remains out” on whether consumers will soon prefer to “self-serve via robo-advice and AI for their mortgages and other financial needs”, but said that he was “reasonably certain that the 50 per cent of home loan customers presently using the services of a mortgage broker might disagree”.

He concluded that while robo-advice would be best suited to "vanilla loans", these are now “few and far between in a world where no two clients are ever the same and where our legislative obligations require us to thoroughly assess each client’s individual needs and circumstances today and into the foreseeable future”.

“The real issue I have with robo-advice and AI is the absence of any qualitative inputs, again within their algorithms, which can foster subjective bias and, consequently, an assessment that is unfair.... [M]uch is learned from a combination of both closed and open questioning — the latter of which remains a significant shortcoming within robo-advice and AI.”

He added: “[W]e who still have an inquisitive mind really need to use it today more than ever before.”

[Related: ‘We're many years away from robo-advice’, says brokerage head]