As opportunities to use artificial intelligence (AI)-powered technologies within aging services communities continue to expand, LeadingAge invited Ethan Mollick, a professor at the Wharton School at University of Pennsylvania and one of the world’s premier voices helping organizations understand how to thrive alongside AI, to share his insights on leading in the age of AI.

I left his keynote presentation on Monday, October 26, thinking about how much of the work ahead belongs to LeadingAge members. Leaders can give staff access to AI, offer training, and encourage experimentation. But they still have to decide what they want to change in their organizations and how those changes will benefit the people they serve.
In his closing remarks, the professor urged aging services leaders to actively participate in shaping how AI is used, rather than leave those decisions to the companies building AI tools. I believe our field has an important contribution to make. Our provider and business partner members understand the daily realities of supporting older adults, working with families, and leading care teams. That experience should guide what the sector asks and expects of this technology; it should inform judgments on how and whether it is helping.

Reconsidering What an Organization Can Accomplish

Mollick’s talk focused on frontier, or highly advanced, AI models, accessed through tools such as ChatGPT and Claude, that enable users to analyze information, explore ideas, and carry out increasingly complex assignments. This was a different conversation from evaluating a specialized product with AI built into it, such as a fall detection system. These broader tools give leaders and staff opportunities to rethink work across an organization.
Last summer, I explored Mollick’s Leadership, Lab, and Crowd framework wherein leaders establish direction, staff discover useful applications, and a dedicated group tests ideas and helps turn them into shared practices. That framework remains useful. But getting better at using AI leaves an important, big-picture question unanswered: What do leaders want their organizations to become capable of doing?

I believe leadership teams need to spend more time on that question. Otherwise, they risk accumulating individual successes without making deliberate choices about where they should lead, what they should prioritize, and how those decisions establish their organization’s direction.

Reshaping Processes and Defining Results

One distinction Mollick made offers a useful starting point. When the purpose of work is to produce an output, such as a report, analysis, or presentation, reconsidering the process to get that end result makes sense. For a more complex task, such as doing an analysis that will inform a decision, AI’s output may offer an approach that looks very different from what’s been done in the past. Leaders need to define what a useful, trustworthy result looks like, establish boundaries, and give teams room to test a different way of achieving it.

For example, before asking AI to help prepare a monthly report, one could ask what decisions that report supports. What does the leadership team need to understand? What information would allow someone to act sooner? Starting with those questions may lead to a more valuable final product than simply producing the same document faster. Mollick’s closing challenge, to consider what organizations should stop doing, follows naturally from this idea. Retiring steps that no longer serve a purpose is a critical part of AI adoption.

Balancing the Value of Person-to-Person Interactions with AI Efficiencies

Mollick also recognized that some work has value because people—not machines—participate in it. A conversation with a resident can build trust and reveal a concern that a questionnaire would miss. Working through a problem with a newer employee helps develop judgment. A team discussion can uncover disagreements that need attention before a decision will succeed. It is important to understand what these activities accomplish before deciding that a quicker result is a better one.

This is where our field’s experience matters. Aging services providers understand the importance of relationships and the consequences of overlooking someone’s needs. Older adults, families, and staff should help determine where AI improves an experience and where human involvement needs to remain central. Their input belongs at the beginning of an experiment, while there is still room to change the approach.

Beyond Speed: Thinking Big Around AI’s Ability to Achieve Larger Goals

Mollick’s invitation to pursue more ambitious goals is the part I hope members carry into planning conversations. Aging services faces big problems: a workforce that is stretched thin, growing demand for support, and too many older adults who cannot find or afford the needed services. These challenges should help set ambitions for AI. Could it help extend the reach of the sector’s workforce, make services more affordable, or support more people as they age in place? Could it help in developing approaches that current staffing and operating models have made difficult to pursue?

AI alone will not solve these problems. But, as widespread AI use is still in its infancy, now is the time to test whether and how AI can change what is possible. That gives experimentation a purpose beyond completing a task faster. Progress can be judged by whether more people receive support, families have an easier time finding help, or staff have greater capacity to meet people’s needs. Those are the outcomes that should guide leaders’ efforts.

An Invitation to CAST Members

Leaders need firsthand experience to make these choices well. Mollick was direct about the importance of learning to use the tools. Judgment that managers and C-suite teams need as decisionmakers cannot be gained through demonstrations or updates from a technology team. Hands-on, direct work with an approved tool on a real planning or operational question helps leaders understand where it is useful, where it falls short, and how much one’s own knowledge contributes to the result. It also puts leaders and managers in a better position to support staff who are learning alongside them.

Through CAST, LeadingAge can help members share that learning and build on one another’s efforts. We all should be as interested in the approaches that disappointed us as in the ones that worked. Both can help the field make better decisions.

I would like every leadership team coming out of the Annual Meeting to choose a significant problem it has struggled to address and ask whether AI opens up a different approach. Set an ambitious goal, involve the people affected, and give staff time to test a concrete first step. Mollick’s closing challenge was an invitation to help shape the future. For aging services, that means putting these tools to work on the problems that matter most to the people LeadingAge members serve.