Should you really use AI tools to manage your money?
It's one of the biggest topics of our time: artificial intelligence (AI).
Touted as being a productivity and technological gamechanger, it has the capacity to influence, even reshape, virtually every aspect of our lives, including our relationship with money.
I asked AI assistant Google Gemini 'What are the pros and cons of using yourself or another AI platform to manage money?'.
The benefits of using AI to manage your money
It promptly returned the following pros:
• automation and efficiency
• speed and data processing
• personalization (to a degree) - (note the US spelling)
• objectivity and reduced bias
• predictive analytics
• cost-effectiveness
• accessibility
• compliance
The downsides of using AI to manage your money
Interestingly, it outlined a larger number of risks (nine risks versus eight benefits):
• lack of human understanding and emotional intelligence
• limited contextual understanding
• 'black box' problem and lack of transparency
• reliance on data quality
• security and privacy risks
• over-reliance and lack of human oversight
• inability to handle complex or unique situations
• regulatory and ethical concerns
• technological limitations and obsolescence
Let's explore these points in more detail, as well as examine other risks that AI itself may not even recognise as being present.
Data accuracy
AI is like a vacuum cleaner - it hoovers up information from all over the internet and then sifts through the material collected.
However, the technology is only as reliable as the data available to it.
There is no shortage of unreliable information online today, including:
• opinions presented as facts
• political propaganda
• marketing spin
• deliberate misinformation and disinformation
• scams, fraud and deepfakes
• outdated information
Using AI does not remove the need to distil fact from fiction.
Indeed, it becomes even more difficult based on the next point.
Lack of transparency
Regardless of which platform you use, AI typically doesn't demonstrate how it arrives at its decisions or recommendations - the sources relied upon and how it weighed them up to steer one way over another.
This lack of transparency limits our ability to fact-check outputs, identify gaps in the decision-making process, and trust the accuracy and completeness of its recommendations.
Product placement
While product placement is nothing new, there is growing debate about how it relates to AI.
Meta founder Mark Zuckerberg recently suggested that advertising will soon use AI to become fully autonomous.
In a financial context, this presents the potential for products and services to be placed (even paid for) yet appear as organically generated recommendations, and potentially flout Australian laws designed to stop vested interests pushing particular financial products.
Input bias
Every AI platform has evolved within certain parameters. These intrinsic biases continue to shape how it operates and learns over time.
The Americanisation of the technology (as previously highlighted in Gemini's default spelling) is one obvious example.
Gender is another, given virtually all the founders (at least of the highest profile AI technologies) are male.
Race, age, sexuality, socioeconomic status, education, profession, health, disability are all potential biases, and more.
AI tools themselves cannot recognise how these biases skew their collation, analysis and presentation of information.
Ethical concerns
Beyond formative bias, there are other ethical risks associated with AI usage.
The most obvious is - are you doing someone out of a job by relying on a machine? Then there are its environmental impacts.
Are you really a sustainable investor if you use AI, given the enormous amounts of energy required to power its servers?
Additionally, there are concerns about AI becoming nefarious.
Think The Terminator movies where machines become self-aware and turn on humanity, and authoritarian countries using AI to wage cyberwarfare.
Untailored recommendations
Regular AI users will be familiar with its (over)generalisations.
The more complex or nuanced an issue, the less capable it is of providing tailored, meaningful suggestions.
Part of this stems from a lack of contextual awareness.
No amount of typing or dictating can provide it with a full understanding of your personal context. Its outputs also depend on what exactly you ask of it.
The fewer specifics you request of it and constraints you put on its search parameters, the less detail it will return.
Privacy concerns
Providing detail and specifics to help AI tailor its recommendations creates a new risk for yourself - privacy.
You also provide identity data simply by using AI, such as your IP address and search history.
Your data, both individualised and collated, can subsequently be accessed by others - both legitimate companies and regulatory bodies, as well as scammers, foreign governments and organised crime.
Unemotional context
Perhaps the greatest limitation of AI is its lack of humanity.
Specifically, the lack of emotional intelligence.
It cannot understand how emotions factor into our financial decisions and desires: from the influences of past trauma to current circumstances, hopes for the future, gut instincts or personal preferences.
Economists struggle with this, relying on data to make predictions which overlooks, for example, how fear drives stockmarkets - herd mentality, panic selling, fear of missing out (FOMO). Computers, by their very nature, are even more susceptible to this.
Cost vs value
Ultimately, like everything money related, you get what you pay for. Free, basic versions of AI give you basic functionality.
Paid-for versions will deliver enhanced features and, likely, superior search options.
Yet they still have substantial limitations.
In Gemini's own words: "AI is a powerful tool to assist with money management, but it should not fully replace human oversight and expertise, especially for complex financial planning."
Get stories like this in our newsletters.
