The annual ASA conference in San Francisco this past March included a panel that addressed the question: “Is AI Age Friendly?” In addition to four in-person panelists (of which I was one), the session started with a video presentation from a fifth participant who provided a short introduction to the topic, identifying some of the potential uses of AI to improve the lives of older adults but also noting some of AI’s potential problems:
As an observant viewer may have surmised, this attractive and articulate presenter is not human. In fact, prior to the session, one of the panelists, Davis Park, had logged on to Chat GPT and given it the following prompt: “create a script for a narrated video under 2 minutes about how Generative AI can help older adults, and outline some of the things we should be careful about.” He briefly edited the resulting script, then submitted it to heygen.com, an AI-powered platform that allows users to generate high-quality videos featuring realistic human speakers. According to Davis, the entire process required a total of about 20 minutes.
Following this introduction, which demonstrated the power as well as some of the limits of the technology, the four of us humans shared our views of this question. Here is our summary:
What’s the Big Deal About Generative AI?
Even though the field of AI is more than 50 years old, Generative AI, introduced just two years ago, represents perhaps the first major paradigm shift since computers were invented in how humans interact with them. Traditionally software has been designed to follow user commands to carry out a specific action (e.g., perform a calculation, retrieve data) based upon predefined rules or logic. Generative AI goes beyond this by autonomously generating new content. Rather than being programmed in advance, Generative AI uses learned patterns from vast datasets (known as large language models, or LLMs) to write essays, generate artwork, compose music, and even create code in response to a simple conversational query.
And because the power of Generative AI is continually being improved, many experts have predicted that this technology will, in a relatively short time, achieve what is known as Artificial General Intelligence (AGI), which means it will essentially equal the cognitive flexibility and problem-solving skills of human beings.
‘Early trials of AI have shown success in rapidly and accurately interpreting medical test results.’
Because Generative AI, even in its current form, is so useful and so easy to use, it has already been integrated into many existing software programs and implemented for a wide range of business applications, including marketing and advertising, customer support, software development, and finance and banking.
Not surprisingly, there has also been a lively interest in Generative AI among providers of services to older adults, who are excited about the technology’s potential to improve the lives of older people by improving the delivery of healthcare and accelerating the development of new medical treatments, improving the effectiveness of services for older adults, and empowering them with new tools for managing their lives.
Some Promising Uses
Generative AI already has demonstrated its potential in a number of different areas to improve the quality of healthcare. For example, early trials of AI have shown success in rapidly and accurately interpreting medical test results, in some cases identifying patterns of symptoms that are too subtle for humans to identify.
Another experiment found that doctors assisted by an AI chatbot did better than a control group without AI in making decisions about how to treat complex medical cases. Generative AI also is accelerating the process of drug discovery by identifying promising drug candidates, as well as finding new uses for existing drugs. And on a practical level, Generative AI is saving physicians time by automating the process of preparing notes from clinical sessions with patients and drafting responses to patients’ email messages.
Generative AI also is being used to improve products and services designed for older adults. For older adults with severe hearing loss, a company called Xander has developed a pair of smart glasses equipped with microphones that use AI to capture speech and project a transcript of what is said that is visible only to the wearer. For people who are blind, Envision has created AI-powered glasses that can read text aloud, detect colors, describe a scene, or recognize a face.
Monitoring systems such as Care Predict have incorporated AI to spot changes in the patterns of activity of older adults living alone that may indicate an increased propensity to fall or other health risks. AI-based services like Lumosity offer AI-generated mind-training games aimed at improving memory and cognitive function. Chatbots like Elliq and care.com are designed to act as AI companions that can combat the isolation of older adults by engaging them in conversations on topics of interest to them, as well as providing timely reminders about taking medications or engaging in healthy physical activity.
Some Caveats
Given how recent Generative AI is, this flurry of activity is impressive. Chatbots have already been used by more than 100 million people, making Generative AI the fastest growing app of all time. But these are still early days, and there are reasons for caution, if not skepticism about where these new capabilities may be taking us, particularly when it comes to their use with older adults. It is always wise to remember that every technology can be a two-edged sword—it could come with unintended consequences that can cause harm as well as bringing benefits.
Here are a few caveats to keep in mind:
First, Generative AI programs create responses based upon what they have learned from extremely large sets of existing content (LLMs). Unfortunately, ageist messages are widespread in our culture and online, with countervailing content often hard to find. To the extent that the input side of the AI learning space reflects a prevalence of ageist stereotypes and negative points of view, Generative AI is likely to replicate biased messaging in its outputs. While users can create prompts to correct for bias, there is a real possibility that, as the general public engages with AI tools over millions of interactions, the AI programs will magnify negative bias against older adults.
There is a real possibility that AI programs may magnify negative bias against older adults.
Second, AI tools are making it more possible to replace in-person care with digitized services. Over time, this could result in fewer care providers and more social robots, less human touch and more technology to care for older adults. If AI provides medical care, social interaction, entertainment and other services without the need for in-person staffing, it could lead to reduced employment in senior care and a further reduction in the human capital dedicated to caring for older adults.
Third, AI may serve as a great accelerator of predatory behavior against older adults. Seniors are already disproportionately targeted with scams, fraud, and misinformation: Calls from people pretending to be a relative in trouble, offers that are too good to be true, password phishing emails. As AI enables more sophisticated, personalized, and lifelike impersonations of people and circumstances, attacks on the security and finances of older adults will become more effective, further increasing the risks for older people who participate in the digital world.
Finally—and perhaps fortunately—the pace of change may be considerably slower than AI enthusiasts are predicting. There remain significant obstacles to full-scale implementation of AI, including inconsistent factual reliability (“hallucinations”), problems with copyright and intellectual property of the source materials used to train AIs, and the potential for regulatory restrictions.
It’s worth remembering that other technology breakthroughs in recent decades—such as self-driving cars, social robots, and virtual reality environments—that had been touted as “game-changers” for older adults have failed to be widely deployed. In general, older adults tend to be “late adopters” of new technologies, and AI may be another example of this pattern. For those who are hoping for dramatic improvements in senior care and services from AI, some timeline caution is in order. This slower pace could provide the opportunity to be thoughtful, and a bit cautious, about how potential AI applications are developed and deployed.
Learning More About AI and Aging
For those interested in learning more about AI and Aging, here are a number of resources to explore:
- Senior Planet has a Virtual AI Resource Center on its website that includes a series of free online classes about AI for older adults.
- AARP’s website includes a section in English and Spanish that discusses “The Possibilities and Perils of Artificial Intelligence.”
Note: The humans who participated in this ASA conference panel on which this blog post is based were: Tom Kamber, executive director, OATS from AARP (moderator); Eric Levitan, CEO, Vivo; Davis Park, senior program director, Los Angeles Digital Equity Action League; and me, Richard Adler, distinguished fellow at the Institute for the Future and chair of Age-Friendly Cupertino.
For more in-depth AI coverage, watch for our Summer 2025 issue of Generations Journal, Artificial Intelligence & Aging, June 25, 2025.
Photo credit: Shutterstock AI Generator
Photo caption: AI-generated image of older adults with family.
