Artificial intelligence (AI) is reshaping industries across the globe, and survey research is no exception. Recent scholarship—including political scientist Sean J. Westwood’s 2025 article in Proceedings of the National Academy of Sciences—raises concerns about AI as a potential threat to traditional survey methodology. At NORC at the University of Chicago, we take a different view: the risk lies not in AI itself, but in its irresponsible use.

Responsible AI is the intentional, fit-for-purpose integration of AI across the survey lifecycle, guided by established survey process and Total Survey Error (TSE) frameworks. A survey process framework outlines the end-to-end process of the survey lifecycle. A TSE framework identifies, measures, and minimizes potential sources of error to preserve survey accuracy. The central risk in using AI is allowing it to dictate decisions.

AI is a general-purpose technology and should not be in the driver’s seat determining when, where, and how it is used. Decisions to deploy AI should not be based solely on AI’s capabilities. Survey methodologists and statisticians must lead these decisions, defining appropriate use cases, levels of human oversight, and evaluation criteria. These decisions should be grounded in established survey frameworks.

At NORC, we use the survey process and TSE frameworks together to provide a structured approach for identifying where—and how—AI can responsibly enhance survey design and data collection while managing risk and preserving quality.

1. The survey process framework enables a systematic assessment of pain points across the survey lifecycle, from study design through data collection, processing, and dissemination. This perspective helps identify where AI can alleviate bottlenecks, address resource constraints, or strengthen execution.
2. The TSE framework complements this approach by explicitly linking AI applications to the error sources they are intended to reduce. It also clarifies trade-offs between error reduction and cost. Critically, the TSE framework provides a principled way to assess new risks introduced by AI, including algorithmic bias, model drift, and other unintended consequences.

Grounding AI use in survey frameworks shifts the focus in two important ways: from AI capabilities to survey needs, and from technological possibility to methodological necessity. Surveys are not simply another domain to showcase AI capabilities. Rather, AI becomes one tool among many to enhance survey quality, efficiency, and rigor.

Responsible AI must be fit for purpose. An AI application may be well-suited to a specific task in one study but inappropriate in another. The key question is not what AI can do, but what the survey requires.

This emphasis on intentionality underscores the importance of human judgment and accountability. Responsible AI also means maintaining quality: AI outputs and impacts must be empirically evaluated against established survey quality criteria, including validity, reliability, and representation.

Finally, responsible AI requires transparency. Organizations should clearly document which AI tools, models, and platforms are used, for what purposes, how models are trained or fine-tuned, and where human oversight is applied. This transparency supports understanding, accountability, and—where possible—reproducibility.

By anchoring AI adoption in established survey frameworks, we can move beyond experimentation toward disciplined, responsible innovation. This approach ensures that AI strengthens—rather than undermines—the core objectives of survey research: producing high-quality, reliable, and trustworthy data.