As the financial landscape evolves, more individuals like Dan, a 41-year-old software engineer from Florida, are exploring innovative ways to manage their retirement savings. Faced with the challenge of investing extra cash within his Individual Retirement Account, Dan sought guidance from ChatGPT.
Initially, Dan aimed to find a single investment option to enhance his current portfolio, which comprised a substantial real estate mortgage REIT and a fund specializing in dividend stocks. However, after engaging with the AI, he restructured his entire $200,000 portfolio following the chatbot’s insightful recommendations.
When Dan input details about his holdings and his available cash, AI suggested a better distribution of his investments. It recommended allocating 80% into a broad market equity index tracker while placing the remaining 20% into a bond ETF—a move that could remain stable for decades.
“I found the AI’s advice to be prudent and sensible,” Dan explains, echoing sentiments shared by other participants in this narrative who opted to use only their first names. While Dan believed he might have arrived at a similar conclusion independently, the interaction with the chatbot gave him the confidence to “commit to and actually execute” his strategy.
Dan’s experience reflects a growing trend among countless individuals across the globe who are leveraging generative AI for retirement planning. A recent study commissioned by Lloyds Banking Group revealed that over half of British adults are utilizing AI platforms for financial guidance.
Furthermore, as many as one-third of users are consulting these tools weekly for financial information or advice, with OpenAI’s ChatGPT leading as the most frequently used platform, followed by Google’s Gemini.
This rapid adoption has also stirred concerns in financial markets, as wealth managers and brokers experienced significant drops in their stock prices after a US-based fintech introduced an AI-driven tool designed to personalize investment strategies for financial advisers.
There are apprehensions regarding how AI might disrupt the traditional wealth management industry. Jason Wenk, CEO of Altruist—a company that developed one such innovative tool—highlighted that it makes “average advice much harder to justify.”
Nevertheless, experts emphasize the importance of understanding AI’s limitations and verifying its suggestions before taking action.
Will, a 32-year-old strategy consultant, had his own brush with the intricacies of using ChatGPT for financial insight. As he relocated from the US back to London, he sought to navigate the complexities surrounding pension liquidations and British tax implications. ChatGPT informed him about a new policy introduced last year that purportedly allowed tax-free withdrawals. However, upon investigation, Will discovered that this was inaccurate.
“The AI lacks the depth needed for detailed inquiries, as it struggles to connect real-world concepts with user-specific situations,” Will remarks.
Mathematical accuracy also remains an area for improvement. Although advanced models are evolving to utilize external tools, such as code interpreters and calculators, to bolster their calculations, inherent errors still occur. These models primarily generate predictions based on statistical patterns derived from their training data rather than performing precise calculations.
A report by the Financial Conduct Authority (FCA) issued in December raised concerns about AI “hallucinations,” wherein AI generates plausible-sounding yet inaccurate information. The FCA recommended that AI users request sources to verify the information against non-AI-generated content.
For those utilizing AI in retirement planning, whether estimating savings needs or determining safe withdrawal amounts, knowing the right questions to ask is crucial. “The quality of the output is directly linked to the inputs; generic or outdated assumptions can arise if specificity is lacking,” notes Paul, a 26-year-old finance employee in Ireland. He found it particularly beneficial for translating broad advice into measurable figures: “If an adviser recommends a specific strategy, I can easily model what that entails over 25 or 30 years. This transparency reduces the ‘black box’ effect and clarifies the value of my investment.”
When we surveyed FT Money readers about AI usage for pension advice, numerous individuals shared their positive experiences. Many expressed the sentiment that AI could eliminate the need for traditional financial advisers, while others with more intricate financial matters turned to technology to minimize the time required with their advisers and to double-check their calculations instead of entirely replacing them.
John Bilton, global head of strategic research at JPMorgan Asset Management, stated that chatbots cannot substitute genuine financial advice. For individuals lacking financial knowledge, chatbots merely provide information in a more efficient format that may still be incomprehensible. “This presents a significant risk,” he warns.
Having spent over 30 years in the financial markets, Bilton prefers consulting an adviser for his pension decisions, acknowledging, “I know what I don’t know.” He expresses concern that the convenience of AI might lead individuals to misinterpret their understanding of financial matters.
Using AI as an investment tool rather than a data tool can inadvertently exacerbate behavioral biases, such as holding excessive cash reserves or trading too frequently. However, the ramifications of AI disruption in finance are still unfolding. In January, the FCA initiated a review led by Sheldon Mills, the former FCA consumer chief, to investigate AI’s future impact on markets, firms, regulators, and retail investors.
In November, consumer research group Which? evaluated six popular generative AI platforms, posing 40 personal finance-related questions and scoring them based on response quality. They found that Perplexity received the highest score, with Google’s Gemini AI following closely. In contrast, Meta’s responses were rated the lowest, lacking detail and specificity.
Andy Laughlin, a principal researcher at Which?, noted that while many platforms produced similar outputs, Meta appeared “rudimentary” and “generic.” Both Gemini and Perplexity excelled in accuracy, positioning ChatGPT as only average. Laughlin pointed out that the AI’s inclination to oversimplify complex topics could obscure their nuanced details, often leading to misinterpretations, especially regarding regulations affecting users in Scotland or Northern Ireland.
Meta opted not to comment on these findings.
Gary Smith, a partner at Evelyn Partners, approaches AI from an efficiency and productivity perspective but refrains from relying on it for advice due to the potential for errors. He highlighted the growing intricacies in pension planning tied to new legislation, asserting the importance of personalized guidance specific to individual circumstances.
“Pensions comprise numerous regulations and product complexities, making accurate AI-generated responses nearly impossible without a thorough understanding of the input,” Smith notes.
Despite its limitations, financial experts acknowledge that AI tools can effectively organize data, suggest ideas, provide alternate viewpoints, and facilitate comparisons of financial products. They can also assist users in crafting questions prior to consulting a financial planner.
Many FT Money readers have found AI beneficial for comprehensive retirement planning, seamlessly blending their financial considerations with personal pursuits.
Greg, a 51-year-old cloud consultant from Northamptonshire, utilizes ChatGPT to devise a retirement roadmap encompassing activities such as language learning and marathon running across different destinations. He employs AI to outline annual goals and determine the necessary pension withdrawals to sustain this lifestyle. Additionally, he values the AI’s ability to validate his plans, revealing previously overlooked aspects like tax-free allowances.
In response to the AI surge, pension providers are racing to keep pace by developing their own generative AI capabilities. These aim to circumvent the pitfalls of mainstream options by guiding users on critical questions and ensuring adherence to the latest regulations.
Scottish Widows, part of Lloyds Banking Group, plans to launch “InvestAI” next month, touting its superiority due to its personalized touch, as it will leverage customer-specific information. Investments director Manuel Pardavila-Gonzalez, overseeing the project, noted, “It’s easy for users to overlook vital information,” adding that Lloyds’s tools will recommend follow-up questions tailored to individual circumstances.
Conversely, other companies trail behind: the UK’s largest defined contribution workplace pension provider, National Employment Savings Trust, is not yet preparing to introduce an AI tool but acknowledges the potential for future AI applications to enhance member customizations and experiences, ultimately leading to better outcomes.
Utilizing tools developed by authorized firms regulated by the Financial Conduct Authority (FCA) offers consumers a significant advantage, ensuring a commitment to client care absent in tools from major technology firms.
However, the precedent of tech integration in investment processes has yielded mixed results. The advent of “robo-advisers” in the UK during the early 2010s faced numerous challenges, with Nutmeg—one of the largest—reporting no profits despite managing £7.4 billion in assets.
After acquiring Nutmeg in 2021, JPMorgan discarded the brand last autumn, opting to leverage JPMorgan’s expertise to provide consumers with superior investment options. Lloyds’ Pardavila-Gonzalez emphasized that prior robo-advisers tended to pressure individuals into hasty investments, contrasting with the bank’s new approach, which prioritizes thoughtful decision-making over swift transactions.
Dan from Florida underscores the value of AI for “financial education and willpower,” expressing appreciation for the ability to engage without the constraints of scheduling traditional meetings. “I feel no pressure to make hasty decisions,” he notes.
