Artificial Intelligence as a Tool for Advocacy and Policy Analysis – Post #4

The assignment: Use ChatGPT or another generative AI product to write a 400-700-word blog post describing how existing policy related to your topic incorporates evidence, and the opportunities to incorporate research and other evidence to innovate and advance policy related to your topic.

According to ChatGPT (personal communication, February 18, 2026),

How Current Nursing Policy Uses Evidence — and Where It Falls Short
Nursing, Professionalism, Health Equity, & Policy Advocacy

In U.S. health policy today, nursing occupies a strange middle ground: it is both indispensable to achieving health equity in underserved areas and simultaneously at risk of being sidelined by regulatory language that undervalues its professional status. This contradiction highlights a broader issue in how evidence is — and isn’t — integrated into policy formation.

At the heart of the recent debate is a policy shift in the U.S. Department of Education’s implementation of Public Law 119-2 (“One Big Beautiful Bill”), which proposes a narrower definition of “professional degree” for federal loan limits under Title IV of the Higher Education Act. Nursing — whether at the master’s (MSN), doctor of nursing practice (DNP), or advanced practice registered nurse (APRN) level — may be excluded from this list, treating nurses as general graduate students rather than professionals. This seemingly arcane bureaucratic definition has big downstream effects: reduced access to adequate financial aid for graduate nursing students, fewer APRNs entering the workforce, and fewer providers in rural and remote areas where they are most needed.

Evidence in Policy: What’s Working

Certain federal policies do incorporate evidence linking nursing workforce capacity to better health outcomes and improved access. The Centers for Medicare & Medicaid Services (CMS) explicitly names health equity as a policy goal in rural and remote communities, and research shows that nurse practitioners in these areas improve access, continuity of care, and patient satisfaction — often with outcomes equal to or better than those of physicians’ care. Evidence from the National Academies of Sciences, Engineering, and Medicine in “The Future of Nursing 2020–2030” underscores that APRNs are prepared to provide primary and preventive care with a focus on social determinants of health, a clear demonstration of evidence impacting high-level policy discourse.

Furthermore, evidence that APRNs comprise a significant share of rural primary care providers — and that they bolster access where physicians are scarce — is supported by workforce data showing APRNs make up a growing proportion of rural clinicians.

Some professional associations (e.g., neonatal nursing advocacy groups) track regulatory changes and translate them into policy briefs or practice recommendations, keeping specialized clinical evidence tied to legislative developments. And scholarly reviews of nursing policy advocacy point to a growing body of literature examining how nursing organizations interact with advocacy processes, even if that work is far from comprehensive.

Policy Gaps: Where Evidence Isn’t Driving Decisions

Despite pockets of evidence-informed policy, several critical gaps remain:

1. Professional Status & Regulatory Definitions: The proposed policy to narrow the definition of “professional degree” essentially treats nursing as not professional in the legal sense. There’s no substantive evidence base presented to justify this shift — no rigorous workforce impact analysis, no data on outcomes tied to educational debt burdens, and no modeling of rural care destabilization. This kind of decision isn’t grounded in empirical research; it’s bureaucratic au courant.
2. Health Equity Framing Without Metrics: Many policies tout “achieving health equity,” but far fewer specify measurable indicators or accountability frameworks. Policies invoking equity or social determinants often lack integration with outcome data, implementation science, or cost-effectiveness research. True progress would require linking policy targets to measurable population health outcomes — for example, maternal mortality rates by geography or race, or APRN workforce distribution metrics over time.
3. Limited Research on Advocacy Effectiveness: While advocacy is recognized as essential, there’s minimal systematic research on what strategies at the organizational or legislative levels effectively shift nursing policy. Existing reviews note a significant gap in scholarly work examining how nursing organizations set policy priorities or influence regulatory outcomes.

Opportunities to Strengthen Evidence Integration

To innovate and advance nursing policy, several evidence priorities emerge:

• Rigorous Impact Analyses: When agencies propose regulatory changes that could reshape workforce pipelines, they should be required to produce transparent impact assessments using workforce data, access metrics, and health outcome modeling. This should be standard practice — not an optional add-on.
• Equity Metrics: If health equity is a policy goal, then policies must be anchored in measurable targets. Health services researchers and implementation scientists should partner with policymakers to define clear indicators and track them longitudinally.
• Cross-Disciplinary Research Translation: Nursing practice research needs to be integrated with economics, health services, and public policy research to influence legislative design, not just clinical standards of care.
• Evidence on Advocacy Practice: We need more rigorous studies on how advocacy efforts by nursing organizations change policy outcomes. Without understanding what works in advocacy, policy recommendations will continue to be shaped more by charter battles than by data-driven strategies.

OpenAI. (2026, February 18). ChatGPT response to prompt about nursing policy and evidence integration [Large language model]. https://chat.openai.com/

Critique of the Chat GPT Post

The professor requests that we evaluate the AI-generated post’s content against the evaluation factors below.

Accuracy, Completeness, Clarity

Overall, the AI-generated content was accurate, but in some areas, it was inappropriately focused. For example, when the breadth of advanced practice is threatened, it focuses on neonatal care, which is relatively narrow and probably a byproduct of my historical interaction with ChatGPT on my account. I believe the summary was clear and would have been clearer with better transitions had we allowed more words. The focus on the use of evidence, I think, added depth to the overall discussion, and ChatGPT came up with some points of advocacy that I felt were strong, e.g., requiring an impact statement related to health equity when the workforce pipeline is threatened.

Implications of AI Technology on this Topic

I believe AI can have a positive overall impact on this topic. Given its national scope, the multitude of professionals affected, the broad effect on various aspects of health equity, and copious data that relate to all the above, I believe that, with the right prompts, AI could have generated a detailed summary addressing multiple aspects of policy, advocacy, and the process of rule change, from numerous diverse angles and perspectives. I find that trying to limit the length of output is where we get into trouble with AI. It condenses information arbitrarily. I would rather have a large production, and then pick through it, find the most outstanding value, and ask it to re-summarize with what I need. Once you help AI focus on what you really want to explore, it becomes an excellent tool for summarizing and creating documents and media that can be easily shared.

Potential Benefits and Risks

The potential benefit is that AI can access and handle vast amounts of information. If you give careful instructions, it can synthesize this information into useful summaries and formats that would probably take a human a week to accomplish. AI learns what is important to you, so if you were focusing on policy, it makes sense to have an account solely based on the policy issues, patterns, strategies, and output that you desire, so that AI learns how to work with you. The risks are incompleteness and errors. When you are not an expert in a topic, it’s hard to evaluate whether something is complete and accurate. I tend to double-check sources and verify their authenticity as a first step. AI likes to invent sources. If the source is wrong, I usually keep asking questions in different ways until it is reciting the sources accurately.

Potential Policy Approaches that Might Be Incorporated to Capitalize on Benefits and Mitigate Against Risks.

As mentioned above, helpful approaches include learning how to best prompt AI to access and output the information you want in the format that you prefer. Learning how to be consistent, handle prompts, and designate certain accounts for a consistent type of work and “givens” seems to help AI stay on track. There should probably be a designated, highly trained human with advanced skills in interfacing with AI, who knows the best AI “machines” and the optimal way to interact with them. Anecdotally, I have found that I learn most efficiently what AI can do by asking for a large output, then refining it, and asking for some information where I can easily spot errors. The significant risk is that you can’t error-check everything, so perhaps an alternate technology could add another layer of error checking. Having more than one set of human eyes on what AI is sending out is essential.

AI is here to stay. I have faith we can improve it to the point where we can trust it more in the future and refine its capabilities to match our needs. I can see the day when one AI machine debates another on policy advocacy, but the implications are a bit scary.