Purpose Before Profit: How AI Can Help People Flourish

For years, the race to build better artificial intelligence has been driven by familiar business priorities: growth, market share and profit. But as AI becomes increasingly embedded in daily life, some scholars and industry leaders argue that its next phase should be shaped by purpose.

For Roshni Raveendhran, an assistant professor at the University of Virginia Darden School of Business who studies the psychology of technology and the future of work, that means building AI that allows people and societies to flourish by supporting three fundamental psychological needs: autonomy, competence and meaningful connection with others.

In this Q&A, Raveendhran explores how purpose-driven AI can foster human and societal flourishing, how people experience these technologies, and why understanding the psychology of human-technology interactions is essential to building AI systems that genuinely serve human needs.

Below is an edited transcript of the conversation.

Q: You study how organizations can leverage AI to generate positive impact for people, organizations and society. Yet recent headlines have been dominated by concerns about AI's harmful effects. How should we think about AI's potential to either help or harm human flourishing?

When AI is primarily profit-driven, the goal becomes maximizing revenue, which often leads to human floundering. When AI is purpose-driven, the focus shifts to improving how well the technology serves its intended purpose, which can enable human flourishing.

The concept of “purpose-driven AI” was introduced by my colleague Nathanael Fast, co-founder and co-director of the Psychology of Technology Institute and Professor at the USC Marshall School of Business at the 9th Annual Psychology of Technology Institute conference, New Directions in Research on the Psychology of Technology. This year’s theme was “Promoting Human Flourishing Through Purpose-Driven AI.”

What stood out to me was how strong and consistent the throughline was across the conference. Regardless of what people are studying, there was a shared acceptance that AI is here, and the focus was on how we can make it more purpose-driven.

It became clear at the conference that researchers, developers, industry practitioners, policymakers, and technology leaders all need a shared framework to think about AI and the idea of purpose-driven AI provided that grounding framework. It really helped orient discussions around a common mission and raised important questions: What does purpose-driven AI actually look like? How can educators, researchers, developers, or policymakers all strive for purpose-driven AI?

At Darden, we think deeply about purpose — purpose-driven leadership, purpose-driven technology, and responsible use of AI. That’s why I was particularly passionate about organizing the conference around this theme. It aligns closely with how we think about responsible technologies and our mission of developing responsible leaders, and how we see our role in shaping the future.

Q: You’ve noted that AI could contribute either to human flourishing or to human floundering. Could you share concrete examples of each outcome?

This was something that was a provocative starting point at the conference. Let’s look at social media. It was originally built to connect people. Over time, it became about engagement, then about maximizing attention, clicks and time spent on the platform. The purpose of the technology kept shifting. That shift has led to a lot of human floundering — doomscrolling, anxiety and other negative outcomes. The research on social media is quite clear: while there are some positive effects, there are also many harmful ones.

In contrast, research by the Neely Center, led by Nate Fast, shows that platforms like LinkedIn or Pinterest are different. Their purpose is more focused — curating professional opportunities or interests — rather than maximizing social engagement for its own sake. LinkedIn didn’t become Instagram. Instagram, on the other hand, began as a photo curation app and evolved into a social media platform with many of the same negative effects we now associate with that category.

Q: How can organizations begin moving in the direction of creating purpose-driven AI — the kind of AI applications that promote individual and societal well-being?

Whether you’re a researcher, educator, technology leader, or a restaurant owner, the key question is the same: What is the purpose of introducing AI in this particular context?

Organizations and leaders need to evaluate AI adoption through that lens and design and implement AI intentionally to improve people’s lives. That means defining metrics that actually align with your purpose. For example, if your stated goal is to increase social connection but your metric is engagement or clicks, there’s a clear mismatch.

With so many AI tools available, it’s tempting to adopt them everywhere. But organizations should ask whether a given tool genuinely helps achieve their goals or whether it disrupts workflows and undermines human capabilities.

Purpose-driven AI requires ongoing evaluation — tracking both benefits and harms and building checkpoints into the process. At every stage, organizations need systems in place to ensure that AI remains aligned with their core purpose.

Q: Your research explores the psychology of technology. What insights do we gain by examining the psychological mechanisms underlying human–AI interactions?

If you understand the underlying psychology of how people experience and work with AI, you can better anticipate its effects across different contexts. One core psychological dynamic is that people expect AI not to judge them. Compared to interactions with humans, people assume AI will not evaluate and judge them.

This lack of perceived judgment means that when people interact with AI, they experience fewer social-evaluative concerns. That single psychological expectation can manifest in vastly different ways depending on the context.

I found that, for example, in a behavior tracking or monitoring context, people tend to be more open and willing to be tracked when AI is doing the tracking. They perceive the environment as informational rather than evaluative. If I take that same underlying psychology and apply it to a different context — such as social support — I see a different outcome. People are often able to feel better after social support from AI (rather than humans) because that support comes with psychological expectations around lower fear of negative judgment.

So while the effects look quite different — greater openness to tracking in one case, improved emotional well-being in another — the underlying psychological driver is the same.

This is why it’s not enough to simply study the surface-level effects of AI. If we only focus on outcomes, we might conclude that AI makes people more open to tracking or that people feel better after talking to AI. But those are different phenomena: one involves behavioral monitoring, the other social support. When you dig deeper, you see they are rooted in the same psychological mechanism. Understanding that mechanism allows us to see how it can manifest across contexts and lead to very different downstream consequences.

Q: Psychologists note that enabling human flourishing requires AI to support autonomy, competence and relatedness. Can you elaborate on what this means in practice?

Autonomy, competence and relatedness are fundamental psychological needs, based on self-determination theory, a highly influential theory in social and organizational psychology. People have a basic need to feel autonomous, to feel competent, and to feel connected to others.

When we study, design, or implement AI, we can’t do so in isolation. Examining its impact on people, and these fundamental needs become critically important. If a technology undermines people’s sense of competence, autonomy, or connection, it will have negative effects.

For example, consider AI companions designed to keep people engaged all day. If the result is that people no longer want to interact with other humans, that erodes relatedness. Or take writing tools: if ChatGPT writes everything for you, it can undermine your sense of competence and autonomy by taking away the opportunity to think for yourself, then that’s a problem.

Given the scale and reach of AI, it is important to be mindful of AI’s impact on these basic psychological needs as it provides a crucial foundation for ensuring human flourishing.