SouthernWorldwide.com – Artificial intelligence (AI) is increasingly being used for a wide array of tasks, from preparing work presentations and assisting with shopping to aiding in scientific research. However, a significant question arises: can AI also be a valuable tool in tackling one of the most complex financial calculations a person will ever face – determining if they can afford to retire?
Data indicates that Americans are already turning to AI for financial guidance. A study by AI company Pearl in September revealed that approximately 20% of individuals use chatbots for this purpose. Furthermore, MissionSquare Research Institute found in a separate study that about half of those who utilize AI at work also employ it for retirement planning, a rate double that of workers who do not use AI.
The need for retirement planning assistance is substantial. Americans now anticipate working four years longer than they ideally wish, primarily due to escalating living costs and insufficient savings. The average retirement account balance for workers stands at $40,000, a figure far below the $1.5 million they estimate is needed for a comfortable retirement.
Adding to these concerns, Social Security, a crucial financial safety net for millions in retirement, faces potential benefit cuts of up to 20% within six years if legislative action is not taken to address its solvency.
How AI Can Assist in Retirement Planning
Given these pressing financial realities, it’s natural to consider using tools like ChatGPT or Claude with queries such as, “Here is my current savings. Will I be able to retire at 65?”
Some financial experts suggest that AI can serve as a useful starting point for addressing basic retirement questions. Luke Delorme, director of financial planning and a Certified Financial Planner at Tableau Wealth, shared his experience. He noted that AI can provide financial planning ideas and even run Monte Carlo simulations to estimate annual spending capabilities, producing valuable output despite not yet being perfect.
A Monte Carlo simulation is a sophisticated mathematical model. It analyzes thousands of potential outcomes for an individual’s retirement portfolio, incorporating various scenarios like market downturns. The simulation then forecasts the probability that a person’s retirement savings will suffice throughout their lifetime.
Delorme believes that such simulations are ideally suited for computer programs and that these AI-driven tools are poised to become increasingly powerful.
Limitations of AI in Retirement Planning
While generative AI may offer benefits for foundational financial planning, experts caution that current large language models are not yet equipped to fully address the intricate layers of retirement planning. These complexities include tax implications and longevity risk, the possibility of outliving one’s savings.
Laurence Kotlikoff, a distinguished economist and retirement expert from Boston University, expressed concerns that AI might inadvertently cause more harm than good when providing retirement advice. He pointed out that these AI applications struggle to grasp the subtleties of Social Security and other retirement-related issues. Moreover, he believes they are often trained on conventional financial planning advice that he considers flawed.
Kotlikoff elaborated that AI’s training data is influenced by Wall Street’s objectives, which prioritize asset management and growth rather than providing truly objective, economically sound advice. He highlighted his own development of a retirement planning tool called MaxiFi.
As an example, AI programs typically estimate retirement savings based on average life expectancy. Kotlikoff argues that retirement planning should instead be based on an individual’s maximum potential lifespan to mitigate the risk of outliving their funds.
Read more: Iranian Destroyers Face Second Attack Wave
He has also observed that AI frequently provides inaccurate projections for Social Security scenarios. This is particularly problematic given the program’s extensive and complex regulations, spanning over 22,000 pages.
Kotlikoff warned that such inaccuracies can lead to fundamentally flawed financial analyses. He acknowledged that AI is currently a highly popular technology, making criticism difficult without appearing outdated or defensive. However, he emphasized his commitment to prioritizing people’s financial security over perceived trendiness.
Insights from AI Models
Andrew Lo, a finance professor at the MIT Sloan School of Management, shared his observations in an MIT publication in April. He noted that AI struggles with tax optimization, lacks an understanding of regulatory nuances, and, unlike human financial advisors, is not bound by legal obligations such as acting in a client’s best interest.
Lo also stressed the importance of critically questioning AI outputs when seeking retirement advice. He recommended prompting AI tools to identify potential errors and to clearly state their assumptions and uncertainties.
To illustrate, consider a hypothetical 50-year-old single woman with an annual income of $70,000. Her retirement savings, primarily invested in S&P 500 index funds, amount to approximately $185,000, which is typical for her age group. She contributes 12% of her income to retirement and is projected to receive about $2,400 per month in Social Security benefits at her full retirement age of 67.
CBS News posed questions to Anthropic’s Claude, OpenAI’s ChatGPT, and Perplexity regarding whether this woman could retire comfortably at 65 and what advice these chatbots would offer. Claude and ChatGPT provided similar responses, suggesting that retirement is possible but would be financially challenging, with a risk of running out of funds under certain conditions. Perplexity was more conservative, indicating that comfortable retirement at 65 is unlikely without significant spending reductions or income increases.
When asked about their underlying assumptions, the AI chatbots indicated they modeled the woman’s lifespan to age 90, whereas a maximum lifespan of 100 is possible. They also noted that they were not modeling precise tax implications. Notably, the AI tools revealed that they did not assess the potentially substantial costs associated with long-term care.
Following this feedback, the chatbots revised some of their initial assessments. Claude, in particular, acknowledged that its original planning horizon was too limited. It consequently adjusted its conclusion from a “tight but doable retirement” to “meaningfully underfunded without course correction.”
A Broader Challenge: Behavioral Barriers
Delorme identified a more significant issue in retirement planning: the pervasive fear of investing among many individuals. This fear can lead to suboptimal decisions, such as holding savings in cash or certificates of deposit (CDs), which often yield returns lower than inflation. Consequently, savings can be eroded over time, increasing the risk of financial depletion in retirement.
Delorme believes AI could potentially help the approximately two-thirds of Americans who do not consult with financial planners to better understand these financial concepts. However, he remains skeptical that AI alone can overcome the deep-seated anxieties many people experience when confronting financial matters.
“It’s much more behavioral than it is a technical lack of knowledge,” Delorme stated. “I don’t know if today that’s going to help people overcome their fears of things, like the fear of investing, which is such a huge obstacle.”
