Young Adults, With an AI Bot, Chat With Their 60-Year-Old Selves

Research progresses on forging closer bonds with our future selves, encouraging behavior helpful to later lives

While the here and now can feel all-consuming, an emerging field of research suggests that thinking about our future self can be psychologically valuable. 

A longitudinal study found that people with a better future-self connection reported higher life satisfaction 10 years later.  Research published in 2024 in the Journal of Research in Psychology makes a case that people with a stronger future self-connection feel their life has more meaning.

It can also help shift behavior. Lab studies have found that people nudged to intently think about their older self are more willing to make choices today that will help them much later on, including saving more for retirement and working out more. It even seems to inspire more ethical behavior.

Yet connecting with our future self typically doesn’t happen organically. Research UCLA Anderson’s Hal Hershfield collaborated on 15 years ago found that when we’re asked to think about our older self, our brain processes that person as a stranger. That insight has spurred research into how to close the gap. Letter writing to our future and past self can help our “future self-continuity.” As can viewing a doctored image of our much older self. (Hershfield’s book, Your Future Self  is an expansive walk through the research and its potential to improve personal outcomes.)

AI and Future You

Now, the arrival of generative artificial intelligence  may help individuals forge a more meaningful link to their future self. 

Hershfield is part of a team of eight collaborators, including MIT’s Pat Pataranutaporn and Pattie Maes, KASIKORN Labs’ Kavin Winson, Auttasak Lapapirojn, Pichayoot Ouppaphan and Monchai Lertsutthiwong, and Harvard’s Peggy Yin, that has launched a new online tool, Future You, that uses Chat GPT 3.5 to create a personalized chatbot that engages participants in text conversation with their 60-year-old self. The researchers in a paper published in 2024 focus on how this tool might improve younger adults’ future-self continuity. Future You is online and free; all participants must agree to be research fodder. 

The key step forward here is that the Future You chatbot runs on a more sophisticated AI technology than the unsatisfying chatbots we’ve all cursed at.  Earlier versions of AI were only able to work off generic information. Think of the customer service chatbot that can only deal with simple generic requests because its technology can only spit out simple generic “help.”

Chat GPT 3.5 harnesses the large language model technology that is capable of processing user-specific data and then creating content that is based on all that personalized information. 

The Future You app first requires users to run through a series of questions. In addition to basic demographic details, participants are asked to share experiences they are proud of and moments that laid them low, career accomplishments, finances, family life and how they hope their (young) life evolves. 

The Future You tech is primed to synthesize that personalized data in a way that builds on prior research about how to foster a stronger future-self connection. Users are also asked to upload a selfie. Once the user has finished with that preliminary step of educating the tool, Future You is ready to have a chat with the user, based on the user’s personal information, alongside an AI-doctored rendition of the selfie that shows the participant at age 60. 

(See a rendering above this article of how the conversation between present self and future self can play out.)

To give Future You a test drive, the researchers sorted more than 340 study participants into four groups. The Future You group had the full experience culminating with a text chat that lasted a minimum of 10 minutes and a maximum of 30 minutes. 

Another group interacted with a more generic chatbot that wasn’t capable of processing personal information. A third group answered the same questionnaire as the Future You group but had no chatbot interactions. A control group merely answered the same pre- and post-survey questionnaires all participants completed, which measured emotional levels such as anxiety and optimism, and participant’s sense of self-esteem and agency.

Participants who went through the full Future You interaction on average reported lower levels of anxiety and negative emotion after the experience than before. They also reported lower levels of feeling unmotivated.

The Future You bot’s ability to impact negative emotions seems to be related to its impact on future-self continuity. Past research has identified three meaningful measures of future-self connection. Participants who engaged with the Future You bot reported higher average levels of overall future self-continuity, a greater feeling of similarity between one’s current and future self, and a more vivid connection to one’s future self. 

While the bot shifted negative emotions, it didn’t have the same power with more positive mindsets, such as participants’ level of agency, optimism and self-esteem. Nor did the Future You bot experience push participants to higher levels of self-reflection and insight.

The researchers are careful to note that, as with all AI,  the possibility for misuse looms — in this instance, suggesting a very negative future — which should inform further research.