AI, Democracy, and Society

The slop economy refers to the degradation of information ecosystems through low-effort, algorithmically optimized content that prioritizes engagement and profit metrics over accuracy and social value. In the age of AI, when systems optimize for engagement, they systematically amplify sensational, low-quality, and misleading content—particularly affecting populations with less institutional trust and weaker media institutions.

AI systems affect democracy through multiple pathways: by embedding developers' worldviews into algorithms, by degrading information quality through recommendation systems, by enabling new forms of surveillance and control, and by distributing technological power unevenly across the Global North and South. The effects are not technologically determined but depend heavily on institutional context.

No. Testing five leading LLMs against a real 420-person survey, synthetic respondents produced technically plausible results but parroted conventional wisdom in harmony with each other, missing every counterintuitive insight that made the human data valuable (Miklian, Hoelscher & Katsos 2026). Synthetic survey data works best as a pre- or post-fieldwork instrument for identifying societal assumptions and conventional wisdoms, not as a substitute for rigorous survey methods.

AI slop is low-quality, mass-produced AI-generated content that saturates digital platforms, and it has also become the dominant insult through which readers police suspected machine writing. Miklian's research covers both senses: the slop economy describes the ad-driven ecosystem of degraded information that billions of users who cannot pay for high-quality content now inhabit, while the AI slop accusation research documents how "that's AI slop" became the standard charge leveled at suspected machine text, accounting for 94 percent of pejorative AI accusations by 2026.

Mostly not. A matched-control test of 421 accused comments against 2,048 controls found that the prose features that statistically distinguish AI text from human text do not predict which human comments get accused (Miklian & Katsos 2026). Accusations track lay heuristics such as polish, formality, and low contraction use, so human writers with formal prose styles are the most exposed to false accusation. The accusation functions as social gatekeeping rather than detection.

It is hardening and institutionalizing. Across 25 million comments from 2023 to 2026, pejorative AI accusations rose more than tenfold, mockery declined from 26 to 7 percent of accusations while structural protest nearly tripled, and ad hoc sneering gave way to formal moderator enforcement of community AI rules (Miklian & Katsos 2026). Anti-AI sentiment has migrated from individual jeers into community-level institutions, and communities built entirely around rejecting AI now produce a disproportionate share of accusation speech.

Miklian (2026) proposes evaluating scholarship on substance rather than process: the scholar must be able to reconstruct the reasoning and defend every claim under sustained scrutiny, regardless of how much AI was involved. Disclosure should describe what AI contributed (brainstorming, synthesis, collaborative reasoning, drafting) rather than a binary yes or no. The appropriate role of AI varies along a gradient of scholarly activities: substantial for information retrieval and synthesis, minimal for interpretive judgment, normative reasoning, and writing as thinking, where the scholar's cognitive engagement is the contribution.

Measurably and fast. AI-associated vocabulary in 15 social science journals rose 151 percent between 2022 and 2025, with the strongest increases in synthesis-heavy journals and the weakest in journals built on original theoretical argument (Miklian 2026). Meanwhile, governance is not keeping up: among 70 top-ranked universities audited across 16 countries, none differentiates AI policy by discipline or epistemological tradition, none addresses AI use in peer review, and none includes a revision or sunset clause.

The new digital divide refers to stratification created not just by access to technology, but by who controls the algorithms, who shapes AI system design, and whose worldviews are embedded in digital systems. Since most AI development occurs in the Global North, the worldviews and biases of Northern developers shape technologies that affect populations worldwide—creating new forms of technological dependence and epistemic inequality.

Democratic regimes may deploy digital technologies for transparency and citizen engagement, while authoritarian regimes use similar technologies for surveillance and control. Hybrid regimes use technology to manage information flows in ways that maintain regime stability. Technology itself is not inherently democratic or authoritarian—its effects depend entirely on the institutional context in which it operates.

Coder worldviews are the assumptions, values, and biases held by software developers, engineers, product managers, and others who design digital systems. These worldviews become embedded in algorithmic architectures—through design choices, feature prioritization, and training data selection—and have downstream political consequences for users, especially marginalized populations.

Miklian's research emphasizes that AI governance challenges are particularly acute in developing countries, which have weaker institutions, less resilient media systems, and less capacity to regulate platform companies. The Global South is disproportionately affected by algorithmic bias, information degradation, and technological dependence, yet has less voice in AI governance and standard-setting.

Deep democracy in technology applies principles of participatory and deliberative democracy to technology design and governance. It emphasizes relational approaches, genuine stakeholder engagement, conflict transformation, and inclusive decision-making—rather than treating technology as a purely technical or market-driven process. It creates space for meaningful participation from those affected by technology.