Today's World Needs Social Scientists

I recently had the chance to speak with a room full of Stanford PhD students about what it means to be a social scientist in the age of AI. I began with a confession: I am not the person closest to the frontier of AI adoption. I have no special expertise in predicting the future. But like many of us, I have been thinking hard about what this moment means for research, for universities, and especially for students just beginning their careers.

My view is simple: there has never been a more exciting time to be a Stanford PhD student in the social sciences.

I don’t mean easy. I don’t mean comfortable. There is real uncertainty, and the anxiety many students feel is understandable. AI is advancing quickly. It is already changing how we code, search, write, analyze, and organize our work and it will likely change much more.

But I also believe the potential for creative, ambitious social science students to do meaningful work has never been greater.

The world needs rigorous scientific insight on social problems more than ever. Poverty and inequality, threats to democracy, regulation of AI, climate change, youth mental health, public health, education, misinformation, the future of work, the challenge of making government work—these are not problems that have somehow become less important because AI exists. On the contrary, the need for credible, reliable, high-quality evidence to guide public and private decisions is growing every day.

At the same time, the power of our tools is increasing dramatically. The firepower available to researchers today, and the firepower that will be available tomorrow, is unlike anything previous generations of social scientists had. That creates an enormous opportunity. But it also raises the stakes in deciding where we direct all of this power.

There are three important reasons why I think the human capital of young, innovative social scientists will continue to be in high demand.

First, we need them to direct our new tools toward the most important social problems. That is not just a technical question. It is not only a math problem. AI can help us reason, model, test, and execute. But deciding what matters requires human judgment. It requires values, intuition, context, and an understanding of institutions and human behavior. It requires knowing not only what we can study, but why the question is worth studying in the first place.

Second, putting social science to work for society requires engaging with people and organizations in the real world. It means building relationships with governments, firms, schools, nonprofits, hospitals, and communities. Robots are good at many things, but they will struggle to replace humans in relationship-dependent tasks like persuading an agency to share data, setting up a randomized trial in classrooms, conducting qualitative fieldwork, filing a FOIA request, testifying before policymakers, or helping a partner adopt a new protocol.

Third, the human systems that are critical for social change will continue to evolve relatively slowly. Computing power, AI capability, and frontier research practices may change very fast. But laws, policies, institutions, organizational routines, disciplinary norms, culture, beliefs, and human behavior generally do not. Many of the most important social problems depend on those slower-moving systems. Changing them requires evidence that people understand, trust, and find relevant. It also requires coordinating change among many humans.

So what should students and early-career faculty in the social sciences invest in today?

Learn and experiment with AI aggressively. The most important work in the coming years will not simply be the same papers written faster. It will be work we could not have written—perhaps could not even have conceived—without these tools. The introduction of computers transformed the frontier of social science. AI may do the same, and likely faster.

Think hard about the why of your work. As our tools become more powerful, the return to asking the right questions will grow. We are likely to see a massive increase in the quantity of research. In that world, the most valuable work will be research whose importance is tangible, whose quality is credible, and whose connection to improving lives is clear.

Invest in relationships. Networks, partnerships, communication, emotional intelligence, and teamwork will remain essential. They will matter for designing research, executing it, interpreting results, and translating evidence into action.

Give yourself a crash course in management. Being an academic has always had something in common with being an entrepreneur. A PhD student is a founder in the garage. An assistant professor is running a small startup. A senior professor is the CEO of a larger enterprise. With agentic AI, our teams are about to get much bigger. The skills required to direct AI agents effectively will look increasingly like strategy, project management, organizational design, and long-range vision.

Build and maintain a reputation for integrity and credibility. In a world of abundant AI-assisted research, decision-makers will face more evidence, more conflicting results, and more uncertainty about what to trust. Being a source that people trust will become even more valuable.

At Stanford Impact Labs, we are thinking actively about how to support this next chapter in social science. We have developed a fast grants program for AI-enabled research to help generate early proof points for how frontier AI tools can accelerate the speed, quality, and relevance of social science in ways that ultimately benefit society. 

This is a moment for experimentation, humility, and ambition. The tools are changing quickly. The problems are urgent. And the work, if we approach it well, is going to get a whole lot more interesting.