Dangers of AI in personal finance

Dangers of AI in personal finance


As Artificial Intelligence improves and begins to be adopted by more and more by firms to improve the goods and services they offer, we as consumers are becoming more accustomed to seeing it around us.

An industry that seems to be very active in adopting Artificial intelligence is the personal finance industry. Due to AI allowing huge amounts of data to be analysed as well as user personalisation and ever improving decision making, AI-based platforms seem to be perfect choice to help users manage their money.

Role of AI in personal finance

The most common way in which AI has helped people manage their money is through personal assistants.By learning from interactions with other users, digital assistants  make it easier to accomplish financial tasks,such as paying bills or managing online bank accounts.Although assistants like these are far from perfect they make a lot of tasks easier and in some cases can give better personal finance recommendations than humans. A good example would be the Bank of america's Assistant known as “Erica”, more on that here.

Another way is through apps that use users data to provide saving and budgeting recommendations.Apps like these take information such as disposable income and current level of spending and recommend ways to save money by allocating their income differently.This idea of using AI to minimise costs will probably be used a lot by firms in the near future to increase their competitiveness.

-Consumer awareness of personal finance apps in the United States in 2016  -via  https://www.statista.com/statistics/637470/familiarity-with-personal-finance-apps-usa/  -

-Consumer awareness of personal finance apps in the United States in 2016

-via https://www.statista.com/statistics/637470/familiarity-with-personal-finance-apps-usa/ -

Benefits of AI-based platforms

Before looking at the dangers of using AI to recommend personal finance decisions I will explore some of the benefits.

One benefit of having AI provide recommendations is to get around the fact humans don't always make completely rational decisions.AI on the other hand just processes the numbers and looks at works best. One example of this would be the concept of “presentism” described by Daniel Gilbert in his book “stumbling on happiness” in which people are unable to imagine themselves in a state unlike the one they are in currently (for example if someone is full now they will be less likely to plan for being hungry in the future).This along with other issues and heuristics such as herding and simple lack of information can lead to humans managing their money in inefficient ways, in this sense AI can help by providing unbiased recommendations.

Another benefit is the extreme amounts of data an AI system can process as well as the constantly improving results (due to the fine adjustment in the weights in the large neural networks as more and more data is processed).Due to these systems being able to do what it would take humans months to do in seconds it can provide significantly more accurate results.

Dangers

However there's are a few issues as we rely more and more on AI for our personal finance.

One issue is consumers becoming too confident in these systems and taking recommendations as gospel.If consumers rely on these systems so much that they begin ignoring their own intuition and rational judgements and the system is flawed there is a huge risk to the consumers of the risk as they trust all their money to a broken system.

Another potential issue is when such systems can make decisions by itself on how to invest the consumers money, this will inevitably be influenced by other systems and can lead to huge instability in the financial markets, such as the “Crash of 2:45” on May 6th 2010 where its possible flawed automated computer systems led to a huge 998.5 points plunge in the Dow Jones (-9%). If the consumers trust these systems to the point they consent to their money being invested, crashes like this could become even more likely, even probable, if they all react to signals in the same way.

Moreover, consumers have no information on how these systems are created or the incentives of the people who build them, the recommendations provided could easily be biased towards certain companies, and we could even see a potential market for firms trying to buy the right for the systems to bias recommendations towards them and other monopoly type issues.

 

Future of AI in Personal Finance

Like any technology there are positives and negatives to its use, however in the case of using AI for personal finance, the positives are huge. The improvements in decision making and data processing can have a huge impact on everyday household spending patterns and the technologies are only going to keep progressing and evolving. However we should never let it become anything more than a recommendation system as there will always be flaws and relying to much on such systems could lead to destructive consequences


Finance, EthicsArturo Dezon