Authored by Yana Kusayeva, Senior Director of Evaluation and Impact
Summary: To better understand how young people are grappling with AI in the workplace, JVS asked them to take and discuss a series of photos. We learned that young workers are already encountering AI in their job search and have a nuanced understanding of AI-related opportunities and risks. The insights from these conversations are changing how we design and deliver training programs.
“AI should reflect the needs of workers rather than set out to replace their labor. Workers must feel valued by their employers and not compete with AI.” – Program Participant
Young people’s voices are often missing in the conversation about AI and its impact on the job market. At JVS, we wanted to better understand how young people are grappling with AI and its impact on their future careers to better inform our program design and delivery amid this time of rapid transformation. So, earlier this year, we invited seven young people enrolled in our dental and medical assistant training programs to first photograph something that reflected how AI might affect their career goals and then photograph something that would make them feel ready to use AI in the workplace. We then came together to discuss what they captured.
We didn’t know exactly what they’d bring back, and the photos and conversations spanned a variety of subjects and themes. In the end, the discussions offered valuable insight into how workforce organizations like JVS can better prepare jobseekers for a rapidly changing labor market.
The project, called “Future, Framed: Work in the Age of AI,” was developed in partnership with MDRC using a Photovoice-inspired methodology. Photovoice—created by researchers Caroline Wang and Mary Ann Burris—uses photography to surface lived experience and community knowledge, particularly on topics that are hard to communicate in words alone. Future, Framed was designed for exactly the kind of conversation we needed to have: one where “AI” means something different to everyone who hears it and where the perspectives of workers themselves are too often overlooked.
How do jobseekers see AI?
Before the first discussion session, most participants had never intentionally used an AI tool. But that didn’t mean they were unaware of AI’s presence in their lives—or their job searches.
One participant photographed a string of rejections on an internal job application system, most of them probably never reviewed by a human recruiter. Participants recognized that AI is already influencing hiring processes, creating an unfair system where employers are encouraged to use AI, while jobseekers are penalized for doing so.
“Precision without Purpose, Replacing the Human Touch. In this photo I am trying to assist, but I am overwhelmed of losing my position to AI.” – Program Participant
Job loss anxiety was the most consistent theme, especially for those in entry-level and service sector roles. One participant photographed a Safeway cashier; another, a Muni train operator—both with the same instinct that these are roles AI may displace. Participants were responding to technologies and workplace changes they could already see affecting the jobs around them.
Who bears the risks of AI displacement?
One of the most striking findings from the discussions was that participants did not primarily see themselves as most at risk from AI displacement. Instead, they named others—older workers, English language learners, people without reliable access to devices—as the ones most likely to be left behind.
Their responses revealed a strong sense of empathy and collective responsibility. Rather than focusing only on their own futures, participants repeatedly considered how technological change might affect others with fewer resources and opportunities. As young people, despite their sense of anxiety about the labor market, they were optimistic about their ability to learn and use digital skills.
One participant’s photo—a conference room screen reading “How To Use AI In The New World”—captured the assumption that training is universally accessible, when for many people, getting to a class at all is a barrier. The photo itself was AI-modified, lending another layer to the discussion.
“As simple as it may seem, it can be difficult for older generations to adapt to new lifestyles. I feel classes or groups that teach older people who aren’t as adapted to new technology can benefit them by getting them caught up to modern and new technology.” – Program Participant
Another participant photographed an iPhone displaying the ChatGPT logo underneath a running kitchen faucet, a deliberate juxtaposition meant to highlight the water and land consumed by AI data centers. They also raised questions about the often-invisible labor and infrastructure behind AI systems—including the workers who help train AI models and resources required to power them—noting that women in rural India are training AI for little pay.
“AI requires the use of data centers that run on drinkable water and require large plots of land. There is a strong effect its prolonged use could have in creating scarcity within capitalism, and consumers who utilize electricity and water in their homes are paying the direct cost of these operations. This image juxtaposes ChatGPT on an iPhone with drinkable water to call attention to the direct cost of running the data machine.” – Program Participant
This kind of systems-level thinking—who owns AI, who trains it, who absorbs its costs—emerged organically from the discussions, showing that young workers are grappling with the impacts of AI not just on their own lives, but on society as a whole.
What could worker-centered AI look like?
The photo that may best capture the project’s spirit is a simple one. It is a mirror with a message written in red marker that looked almost like a protest sign: “AI should reflect worker needs!” The participant’s caption called for AI that supports workers, rather than competes with them, and for employers to be accountable for how they deploy it.
In discussions, participants wanted transparency about when and how AI is being used in hiring and explicit agreements clarifying that AI would handle tasks, not replace people.
“This photo displays a dental assistant (myself) working on a patient. In the back there is a dark figure giving me a thumbs up (AI). The photo represents AI as a resourceful tool assisting me with charting as I am assisting the patient.” – Program Participant
At the same time, they had concrete ideas about how AI could help them with their day-to-day work. Dental and medical assistant participants described specific ways AI could improve their work, including assisting with cavity detection on X-rays, managing inventory, and reducing documentation burdens during patient visits.
The distinction they drew, consistently, was between AI that handles the repetitive and the administrative, and AI that displaces the human relationship at the center of healthcare. Participants were emphatic that patients want to hear from a person, not a chatbot, about their care.
What JVS is doing with what we’ve learned
The Photovoice discussions had a surprising, immediate impact. The facilitated conversations themselves were sufficient to pique interest in AI tools, and participants who arrived with apprehension began experimenting with tools on their own between sessions. Moreover, one participant who had not used AI before the project’s first session started using ChatGPT to tailor their resume and assist in writing cover letters. Between the second and third sessions, they got hired.
The conversations also reinforced something we believe at JVS: Workforce organizations have a responsibility not only to teach people how to use new technologies, but to help them critically evaluate how those technologies are impacting their work. Participants wanted practical AI skills, but they also wanted transparency, accountability, and a voice in how AI is implemented. Moving forward, we are exploring how to embed both AI literacy and facilitated peer conversation capable of shifting attitudes into our training programs.
We’re also rethinking how we use Photovoice itself. More than a program improvement tool, this project showed us it can function as a program design tool, a way to hear from participants early in the process of building or iterating on something new. We are also exploring its potential as an intervention tool for leadership development for JVS alumni and as an ongoing feedback mechanism across programs.
For JVS, the goal has always been to ensure that the people we serve aren’t just trained for the future of work, but help shape it. Future, Framed gave us another tool for doing that—and more evidence that the people closest to the problem are often the ones who can see the path forward most clearly.
Join the conversation: Share your thoughts with us on LinkedIn.
