Policy Theories in Crisis: AI’s Impact
Tackling AI in Policy: From Confusion to Clarity
Imagine this: A policy decision that could save lives is overturned, not because citizens are overwhelmingly against it, but because an artificial intelligence-powered platform bombarded decision-makers with machine-generated comments.
This is not a scene from a science fiction novel, but a real event. In June 2025, Southern California’s Air Quality Management District faced just such a scenario, overturning proposed rules to regulate emissions after an influx of fabricated public comments. The implications for democracy and policy-making are profound.
Understanding the Crisis: Why Conventional Policies Are Dying
The policy-making landscape is undergoing seismic shifts. Traditional theories of agenda-setting, such as Kingdon’s Multiple Streams Framework and Baumgartner and Jones’s Punctuated Equilibrium Theory, have been pillars in the study of policy processes.
These frameworks, however, relied on assumptions now being challenged by technology: that information is authentically human, filtered by credible sources, reflective of genuine public opinion, distinguishable from noise, and anchored in verifiable events.
The rise of AI and large language models (LLMs) has fundamentally changed these assumptions. Political information can now be machine-generated, indistinguishable from human input, bypassing traditional gatekeepers and creating a simulated public sphere where public sentiment may not be genuine.
The New Reality: Five Changes AI Brings
- Unlimited Political Content: Machines can generate information at an unprecedented scale.
- Collapse of Gatekeeping: Traditional filters, like media institutions, struggle to verify and mediate content.
- Contaminated Opinion: AI can skew public opinion measurements, influencing policy based on synthetic data.
- Epistemic Recursion: Training data for AI becomes a self-referential loop that distorts public perception.
- Orchestrated Simulations: AI can simulate social movements, creating phantom crises and policy shifts.
What This Means in Practice
For public health professionals, researchers, and policymakers, these developments demand a rethinking of engagement strategies and the tools used to interpret public sentiment:
- Embrace new technologies for content verification and authentication.
- Develop policies that ensure transparency in the origins of information.
- Engage in continuous education on AI’s influence on public policy and decision-making processes.
Looking Forward: Navigating AI Challenges
The transformation necessitates rethinking current frameworks. Epistemic transparency, signal integrity, and event verifiability must be defined and measured to maintain policy processes that are reliable and reflective of citizen interests.
The policy field needs to evolve by developing the Epistemic Policy Process (EPP) theory, which addresses these new challenges head-on and proposes a framework in which policy processes adapt to technological influences. This theory underscores the importance of resilient, robust policymaking that adapts to the evolving information landscape.
‘Policy theories are not inherently flawed, but when information origins are unclear, applications must shift.’
Conclusion: Embracing the Uncertainty
As we move into this uncertain terrain, public health leaders, policymakers, and civic organizations are encouraged to foster dialogue on embedding transparency and accountability in the policy process.
- How might AI be integrated into local policy-making processes without bias?
- What resources are necessary to develop AI-informed policy frameworks?
- How can your organization advocate for increased transparency in policy information streams?
In the age of AI, the challenge is to ensure our policies remain products of informed decision-making — not simply reflections of the content that machines can create.
