Developed by the MIT Media Lab, the MemPal system was co-designed with older adults, enabling them to shape a future where they can age comfortably and safely at home. Here, we share anecdotes from user testing of MemPal in real-world home environments. 

Memory is of utmost importance and I want to keep my memory for as long as I can. 

– Rochelle, a 78-year-old tester of the MemPal system

The world is undergoing a major demographic shift: the elderly population is rapidly increasing. By 2034, Americans 65+ will outnumber children under 18 for the first time in history. In the US especially, there is a lack of community, affordable caregiving support, nor is there much in terms of technology to support them. Memory loss (whether cognitive decline, dementia, or Alzheimer’s) is an inevitable part of aging. One-third of older adults experience it, placing immense pressure and stress on (often unpaid) caregivers as well. 

Over the past year, researchers in the Fluid Interfaces group at the MIT Media Lab developed MemPal, a wearable voice-based memory assistant, designed to support older adults in maintaining their independence and safety within their own homes, rather than having to move to nursing homes or the home of a family member. 

MemPal started with a clear purpose: to create technology that helps older adults live independently and safely in their own home and improves their quality of life. The needs of older adults are often overlooked due to the challenges in designing such products, the tendency of younger developers to ignore developing for a elderly population, or the belief that the market is too small. As AI technology rapidly improves, we should not ignore its potential to help often-neglected groups. I believe the elderly could benefit the most from these emerging technologies. 

MemPal is a wearable assistant designed to ease the daily challenges faced by those struggling with memory issues. With a gentle, reassuring voice interface, users can simply ask, "Hey Pal, where's my keys?" or "Hey Pal, did I remember to lock the door?" MemPal responds instantly, providing peace of mind without compromising privacy. 

MemPal offers subtle reminders to keep users safe and on track: "Did you remember to turn off the stove?," or "You already took your medicine an hour ago." It can keep loved ones informed as well, passively tracking completed activities in its app and providing summaries to caregivers to help reduce their worry and stress.

For doctors, MemPal can generate a concise report of activities and track aspects of memory decline, offering valuable insights that can help physicians better understand their patients' abilities. This window into everyday behavior could prove crucial in identifying and addressing memory conditions like dementia and Alzheimer's early on.

To enable these features, MemPal uses a camera worn around the neck to observe the world around its user, specifically focusing on the user’s hands. This visual context detection is powered by a multimodal large language model (LLM) system which creates a real-time automated text diary of all actions performed by the individual, which can later be queried for memory support, offering object retrieval assistance, recall of past actions, and proactive safety reminders using a voice-based interface. 

Behind the scenes, MemPal's technology blends advanced computer vision with language understanding, working in parallel to constantly yet unobtrusively watch over the user's environment to offer practical, empathetic support right when it's needed most. 

The research team started with a user interview phase: gaining insights from physicians, older adults, and caregivers through Massachusetts General Hospital (MGH) support groups to understand their pain points and frustration caused by looking for lost objects, forgetting reminders, constant caregiver monitoring, and not having accurate information for diagnosing memory conditions. 

With a plan in place, we presented MemPal at senior homes across Boston, and put up recruitment flyers at churches and libraries across the city. There was significant interest in joining the trial.

Months of development and prototype testing followed to match the needs of users, such as designing a unique room localization algorithm for user-friendly setup of the Mempal system. The final system combined a wearable device with an app. Our key test: determining if older adults found MemPal user-friendly and effective for memory support.

We began a user study to trial the device among 15 older adults within their own homes. These are three of their stories.   

***

[TEST 1 ] 

Our first participant was Sharon, a devoted caretaker for her husband with Alzheimer’s, but suffering from subjective cognitive decline herself. She welcomed us into her home, and kept apologizing for her husband’s behavior, who began confabulating when he had dinner, arguing when he could have dessert.

As she proceeded with testing the MemPal device, she discovered how its natural voice interface could save her time and daily frustration: “Wow, I can just ask it like, ‘Hey Pal, did I make him dinner?’ This would save us so many arguments. It would be really useful for more elderly caregivers who are likely to not trust their own memory.” Sharon's relief was palpable as she explored the device's functionality of finding misplaced objects in her messy house and having proactive safety reminders, which she found surprisingly intuitive and helpful. Just as we were about to call an Uber back, she asked, “Right now I just keep monitoring him with Life360. Can I put myself on the waitlist? When will I have access to one of these MemPal devices?” It was only our first test, so I couldn’t give her an estimate!

[TEST 2 ] 

Another participant was Rochelle, a 78-year-old grandmother with three grandchildren. She lives alone after losing her husband in a hit and run accident 10 years ago. As we set up the user test of MemPal, Rochelle tearfully shared her story with us. As the circumstances and the people in her life change, she fears the idea of losing her memory as well. 

Rochelle, sitting by a big window in her New England apartment, seemed poised and proud as she held up picture albums and gushed about the upcoming graduations and weddings of her three grandchildren. But outside the picture albums, her apartment seemed empty. Rochelle was still reeling from the loss of her husband. “They were 53 wonderful years.” 

Although her kids live nearby, they have full-time jobs and can’t take care of her 24x7—and Rochelle's frustration over losing items or forgetting simple tasks was getting worse. “I often tend to forget my cane and keys when I walk out the door.” When Rochelle used MemPal to help find her phone or her keys, she was surprised at its accuracy and detail. “So specific!” she exclaimed.

Notably, she noticed the critical need for a device that could provide a detailed record of daily activities and health metrics, offering a more comprehensive support system. She started to recount how her sporadic doctor appointments based their diagnosis on subjective answers: “I have lost weight when I eat normally. The doctor asks me why? I don’t remember anything different from my usual habit. So, she tells me to come back every month just to hear the same answer. I think it is necessary for the doctor to know my daily activities within the home [what I’m eating, when I’m eating, what activities I do]. Definitely.” Before we left her home, Rochelle gripped my hand and asked me to stay a little while longer.  “Memory is of utmost importance and I want to keep my memory for as long as I can.”

[TEST 3 ]

The next participant, named Carol, invited us into her home, which was a colorful blend of creativity and chaos, with vibrant paintings and a cluttered table filled with medications, papers, envelopes, and colored pencils. The struggle to manage her medications was evident—the extensive list of medications on her refrigerator was programmed into Med-Minder’s buzzing reminder system. “Can someone turn that off?” she yelled, “It doesn’t even help.”

But when Carol tried MemPal’s safety reminder system, she noticed that it recognized her context, like completing a meal in the kitchen before reminding her to take medication. After letting Carol try the safety reminder feature, she was reminded of an often repeated incident: “I had turned the stove on when cooking, then left my apartment to go to an event. Apparently there was a siren but I’m deaf in my left ear. I saw a fire truck outside the building and realized it was for my apartment. The stove was still on.” After the test, she requested a brochure to share with her friends: “So many people need this. This would be so useful. I hope one day I can have one.” 

These and about a dozen other trials of the Mempal system were performed in people’s own homes in the last month. 

***

Our initial user interviews made it clear that MemPal was a much needed innovation to improve the lives of older adults. But these in home user studies actually proved its effectiveness. The user tests with older adults in their own homes highlighted MemPal’s potential to support memory and cognition and enhance quality of life. The in-home studies also resulted in invaluable feedback for developing the next iteration of the system. Sharon’s caregiving difficulties, Rochelle’s daily struggles, and Carol’s medication management each highlighted the importance of MemPal’s practical support in easing their stress and empowering their lives. 

Unlike my grandparents in India, who are surrounded by family, full-time support, and community, many elderly in the US live alone with minimal assistance and few family members nearby. In the US, after a certain age, it almost feels as if you are a burden on society: “There is no space for me,” some participants would mention. In nursing facilities, people are constantly checking up on you, and you lose the freedom to live life as you once did. 

I felt proud that the elderly with memory-related conditions were not being left out of technology innovation. We involved them in a co-design process to help us imagine a future where they can live comfortably with many of their daily struggles resolved. 

I learned firsthand that older adults still have dreams, a wealth of experience, and a lifetime of memories they wish to preserve as long as possible. 

This work would not have been possible without the contributions of the following people: MIT Media Lab postdoc Samantha Chan, my mentor and close collaborator, and Professor Pattie Maes, PI of the Fluid Interfaces research group, as well as the Fluid group researchers that assisted with user testing and app development—Christine Xu, Scott Ren, Rachel Park, Adan Abu Naaj, Aditya Suri, Nigel Norman, and Arnav Kapur. In addition, I would like to thank Jimmy Day and Cecilia Nafkoor for documenting this journey.