Edition

The Impact of Artificial Intelligence/Large Language Models on Retirement Professionals and Retirees Call for Essays

Background and Purpose

Hardly a day goes by without a feature story about the latest reverberation from the expanding use of Artificial Intelligence/Large Language Models (tools) in a variety of settings. The pace of use (and misuse) of these tools, such as ChatGPT and Bard, has been explosive and shows no sign of abating. Many experts would argue we are on the cusp of a permanent paradigm shift in teaching/academics and overall application of technology.

If and how this shift plays out, there is reason to think that it could impact in significant ways the nature and extent of work currently done by actuaries, financial planners, advisors, attorneys and other retirement professionals. This could be as little as minor assistance with report writing to full delegation of analysis heretofore thought to require deep professional knowledge.

Outside of professional settings, retirees and those nearing retirement themselves could significantly feel the reach of these tools. For example, they may change the method of delivery of investment advice, asset allocation, and general planning. At the same time, though, these tools may create new risks that expose retirees in ways that they have not and could not have been prepared.

With these issues in mind, the Society of Actuaries (SOA) Research Institute’s Aging and Retirement program is interested in an exploration of this topic from a variety of perspectives. The result of this effort is intended to provide a useful resource for readers to become well-grounded on the issues surrounding the impact of these tools in a retirement context and set the stage for future research.

Research Objective

The sponsors are seeking essayists to explore the multifaceted impacts of AI/Large Language Model tools on retirement professionals, retirees and those planning for retirement. The desired end product of this effort is a resource that will aid in the understanding of relevant issues while providing current considerations and potential future dynamics of AI/Large Language Models in this area.

The intended audience for this effort is actuaries and other retirement professionals such as financial planners.

The following are illustrative of the topics and questions that might be addressed in an essay. Respondents to this Call for Essays are not required or expected to address all or even any of these topics/questions. Rather they are encouraged to also consider other question(s) or topic(s) they deem critical to the work done by the Society of Actuaries and its members.

Responses to this Call for Essays should clarify the time horizon under consideration in an essay. The SOA Research Institute prefers that essayist address both current and future professional and retiree considerations of the AI/Large Language Model questions or topics they plan to focus on. Respondents are, however, free to cover only one aspect, but in either case, the responses should clarify what will be covered.

Retirees and Those Planning for Retirement

  • What are the risks and challenges that these tools may present specifically for the aging community? For example, what are risks to individual investments they may hold?
  • What is the best way to educate retirees and those planning for retirement on the uses of these tools?
  • What are ways to address fears that these individuals may feel about the adoption and use of such tools? What are concerns that are warranted for these individuals and which may be dispelled with better information?
  • What are the ways that these tools can best be used to educate and otherwise support individuals in managing their personal finances during their retirement years?
  • How do you make these tools useful for people in retirement if they may end up being the last to use them and the most exposed to abuse by them? How can tool interfaces be improved in this regard? How can their underlying algorithms be improved?
  • What may be risks introduced by using tools with algorithms that may be potentially biased or discriminatory? How might such biases, that may be racial, gender or age-based, specifically impact an aging population?
  • What are ways to prevent fraud and scams directed at retirees that make use of these tools? What are risks posed by such fraud and how can AI itself be used to help educate individuals on it?
  • How might these tools be used to promote better living in retirement? Can they help with more easily living independently, managing bills/payments, monitoring cognitive decline, etc.? Are there ways these tools can provide support to and lessen the adverse consequences associated with memory loss?
  • Can these tools help retirees and those planning for retirement better navigate getting help, such as more efficiently reaching customer service centers?
  • Can these tools help retirees who wish to work part time in retirement or otherwise return to the workforce?

Retirement Professionals

  • What are ways that these tools can be used for preparation of actuarial analyses for retirement plans (defined benefit, defined contribution, etc.)? What are the most beneficial uses for such analyses? Which types of analyses may need to be approached with greater caution? For instance, how do you assess whether outputted information is reliable vs inaccurate/fabricated in this context? Can these tools themselves be used to help make such assessments? Which types of analyses are completely inapplicable or in other ways limited? How do the preceding considerations differ between now and in the future?
  • How would the issues in the previous bullet apply to retirement professionals who advise on retirement planning and prepare individual projections and other analyses?
  • What are uses of these tools for preparing reports that retirement professionals may need to create for any purposes outside of actuarial analyses? How might this impact the current roles and responsibilities of retirement professionals at different career stages? Are there positions that may become obsolete?
  • To what extent can these tools be used to professionally manage retirement investments and portfolios? How can these tools be used to help market such services?
  • How can prompts be written? What needs to be watched out for?
  • How might these tools support plan administration and reduce the need for personalized communication in writing and by phone? How does one avoid a participant getting stuck without an answer?
  • How do employers of retirement professionals feel about these tools? Are they embracing the tools for their employees’ use or are they steering employees away from them? What discomfort may they have?

Again, the list above is not meant to be exhaustive but merely examples of topics that may be examined in an essay.

Rules for Submission of Essays

Timeline

The deadline for submissions is April 9, 2024. Essays will be published in a formal collection as soon as possible.

Special Consideration for Essays:

Given the topic of this call for essays, two categories of essay submissions have been established. Category 1 includes essays written by individuals without help from any of these tools. Category 2 is for essays written primarily or partially by one or more of these tools. Each category of submitted essays will be reviewed separately for publication and award determination. Authors should note which category they are submitting their essays for consideration. For Category 2 essays, the extent that the tool was used should be clearly identified and it would be helpful to describe what prompts were used and why they were used.

Length and Instructions for Submission

Essays must be submitted in English with a desired length of between 500 to 2,500 words. There is no requirement for formal or extensive footnoting.

Author information must be submitted with the essay and include name; credentials or designations (if appropriate); title; organization/company; e-mail address; and phone number. Please provide all author information at the beginning of the essay.

Essays that contain any overt political statements, commercial content, and other inappropriate material will not be accepted. Articles must comply with the SOA's antitrust guidelines.

Please submit your essay via e–mail to Research-AR@soa.org.

Awards

Up to $5,000 in award money has been allocated for this call for essays. The review committee will select the leading essays and determine how to allocate the award money among them. Consideration will be given to creativity, originality and the extent to which an idea might help promote further research or thought in this area. In exchange for award money, selected authors will be required to assign all copyrights in their essays to the Society of Actuaries Research Institute.

Authors are ineligible for awards if an essay is based on an Institute-sponsored and funded research study conducted by the author. However, authors are welcome to submit such essays for publication consideration.

Publication and Presentation

Depending on how many essays are received and the range of the topic areas, a suitable format for electronic publication and dissemination will be selected. Essays may also be presented at an SOA meeting, webcast, or other professional development event.

In addition, other venues for publication or presentation of the ideas outside of the Institute will be considered. It is hoped that publication of the collected essays will further knowledge and stimulate discussion as well as promote future efforts in the area of Artificial Intelligence/Large Language Model tools.

Rights Granted

Please understand that by submitting an essay for consideration, the essay author(s) is granting to the Society of Actuaries Research Institute an unlimited license to print or republish their essay, with proper attribution given to the author(s).

Questions

Please direct any questions regarding this Call for Essays to Research-AR@soa.org.