Your grandma should be using AI. Really.

A staffer to a member of Congress recently asked me, “How much are elderly Americans using AI?”

My rapid answer? Not enough.

The longer answer: AI can meaningfully improve both the quality of life of older Americans and the support systems that surround them—and policymakers are leaving that potential on the table.

As it stands, a large fraction of older Americans are missing out on the AI wave. More than 50 percent of the Silent Generation (1928–1945) report that they have never used AI. The same is true of 39 percent of Boomers (1946–1964). They’ve yet to see if it can help them write an email to their doctor, research a proposed change to the city charter, or create a funny image to send to a friend.

This lack of use is unsurprising given that many older Americans are unfamiliar with AI tools. Half of members of the Silent Generation have heard “hardly anything” about AI. Around a quarter (24%) of Boomers likewise report being in the dark. This lack of familiarity translates into a lack of excitement: the vast majority of older Americans report little to no excitement about AI—64% among Boomers and 77% among the Silent Generation.

Perhaps paradoxically, they are also very concerned about AI. When given the option to select between “not concerned at all” and “very concerned,” hardly any older Americans selected the former, whereas 39% of Boomers and 31% of Silent Generation members flagged a high level of worry.

The explanation for this apparent contradiction is straightforward: most older Americans have never encountered AI use cases tailored to their needs. Prior general-purpose technologies, like electricity, were much easier to explain and demonstrate to Americans of all ages. The flip of a switch enlightened Americans to the potential of the technology to transform just about every aspect of their lives. For some Americans—particularly those who already spend an inordinate amount of time on their phone and computer—the introduction of AI led to a similar experience. For many others, however, there’s a larger barrier to adoption when it comes to discovering the best use cases of AI.

It’s imperative that policymakers and innovators tackle this usage gap. The goal isn’t to bolster the bottom line of one company or another but rather to make sure that Americans of all ages have access to technology that can help them live healthier, happier lives.

Let’s start with an easy one: autonomous vehicles (AVs). When my grandma was no longer able to safely drive, our family had a difficult time keeping her away from the wheel. She did not want to miss out on the freedom and agency associated with being able to go anywhere, at any time (my grandma would never step into someone’s car without glamming up beforehand). It’s a moment of tension many families know well. Yet AVs like Waymo can make that conversation much simpler, offering a safe, reliable, and private experience. Well, if you can access it. As it stands, Waymo and other AV companies operate in just a handful of jurisdictions. Policymakers can and should lower barriers to AVs launching in more places. They may also want to consider means to make rides more affordable for seniors, such as via block grants to local nonprofits and care providers.

Mobility and companionship may seem like separate problems. They share a common policy failure: regulators have yet to realize that inaction is perpetuating a status quo in which older Americans stand to benefit from innovation.

A more complicated case study takes a different form—specifically, the shape of a small, almost lamp-shaped robot. The New York Times recently profiled ElliQ, a robot specifically designed to help address the loneliness epidemic that has spread rapidly across the U.S. Its founder explained that he hopes it behaves like a kind, teenage granddaughter—inquisitive, optimistic, and willing to check in on things like how a user is feeling. It’s not a necessary tool for all older Americans, but it may be a major quality-of-life improvement for others: pilots in New York and Washington found that people interacted with their ElliQ more than 40 times a day, and nine in 10 users reported feeling less lonely. That’s a big deal, especially given that ElliQ and related tools will presumably only improve over time.

Waymo, ElliQ and a growing number of AI tools aim to solve problems that led my grandma to regularly remind me that “growing old isn’t for sissies.” So long as these tools fly under the radar, though, their potential upsides will go unrealized. That’s a preventable future.

It’s also a future legislators can make less likely by encouraging pilots just like those run by Intuition Robotics and others. Now’s the time to launch regulatory sandboxes across the country that allow innovators to deploy tools subject to close monitoring from the relevant state or federal actors. These test beds for innovation can lower barriers to adoption for older Americans while also ensuring that flawed tools are identified quickly.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.