Semantic Network

Interactive semantic network: At what point does a platform’s effort to curb misinformation infringe upon legitimate investigative journalism, and how can policy distinguish between the two?
Copy the full link to view this semantic network. The 11‑character hashtag can also be entered directly into the query bar to recover the network.

Library

Q&A Report

Curbing Misinfo: Where Does Free Press Meet Regulation?

Analysis reveals 9 key thematic connections.

Key Findings

Editorial Labor Devaluation

Platforms' algorithmic enforcement against misinformation after 2016 inadvertently degraded investigative journalism by equating time-intensive, source-dependent reporting with real-time false claims, collapsing epistemic distinctions in content moderation systems. Moderation infrastructures built to respond rapidly to viral falsehoods—such as third-party fact-checking partnerships and automated takedowns—prioritize speed and scale, marginalizing journalistic practices that rely on delayed verification, anonymous sourcing, and institutional accountability, thus redefining rigorous reporting as procedural risk. This shift, crystallized during the post-2016 electoral disinformation crackdowns, exposed how platform governance began treating editorial judgment not as a safeguard against misinformation but as a liability akin to malinformation, particularly affecting local journalists and whistleblowing intermediaries. The non-obvious consequence is that the very actors historically tasked with uncovering institutional deception became structurally vulnerable to the same moderation mechanisms designed to protect public discourse.

Gatekeeping Legitimacy Transfer

The transition from legacy media dominance to platform-mediated news distribution after 2010 transferred gatekeeping authority from newsrooms to tech firms, allowing platform policies framed as neutrality to selectively delegitimize investigative work that challenges powerful actors. As platforms assumed editorial-like roles through recommendation algorithms and community guidelines—especially after the 2020 U.S. election cycle—they began to arbitrate truth claims not by journalistic standards but by reputational risk, resulting in the suppression of investigative reports under broad misinformation labels when they involved classified leaks or corporate wrongdoing. This reconfiguration placed investigative journalists aligned with marginalized communities or adversarial to state power—such as those covering police misconduct or surveillance programs—at disproportionate risk of content removal, revealing how platform policy, under the guise of consistency, replicates state or corporate norms of legitimacy. The underappreciated outcome is that the historical function of journalism as a counter-power has been destabilized not by overt censorship but by the temporal and procedural mismatch between investigative timelines and platform enforcement cycles.

Accountability Evasion Infrastructure

Following the 2018 Cambridge Analytica scandal, platforms formalized misinformation policies that obscured their own role in amplifying deceptive content, framing enforcement as neutral hygiene rather than political judgment, thereby insulating themselves from accountability when those policies hinder investigative journalism. By adopting standardized definitions of misinformation rooted in consensus verification—often tied to government or scientific authorities—platforms avoided responsibility for decisions that disproportionately removed content exposing institutional malfeasance, especially from freelance or nonprofit investigative units operating without institutional backing. This shift toward procedural compliance over epistemic transparency created a system where the removal of investigative material could be automated or outsourced without public justification, effectively transforming content moderation into a mechanism for deflecting scrutiny. The overlooked consequence is that platforms, under pressure to act decisively against misinformation, built infrastructures designed not to resolve truth claims but to minimize liability—rendering investigative journalism a collateral casualty of institutional self-preservation.

Temporal Preemption

Efforts to suppress misinformation cross into suppression of investigative journalism when preemptive content takedowns based on probabilistic risk models displace stories before publication, particularly in cases where journalists rely on platform infrastructure for source coordination and encrypted data transfer, as seen in the throttling of Signal-linked file shares or shadow-banning of draft tweets under AI surveillance systems like Twitter’s Trust & Safety algorithms. The overlooked dynamic is temporal preemption—the platform’s ability to act on anticipated harm before an investigative piece materializes—transforming risk assessment from response to anticipation, which shifts the burden of proof onto the journalist to prove legitimacy in advance, a reversal unseen in traditional media law and rarely acknowledged in transparency reports; this creates a hidden chokepoint at the pre-publication stage, where the speed of algorithmic governance outpaces editorial timelines and deters high-risk reporting by design.

Moderator capture

Platform content moderation during the 2020 U.S. election disproportionately removed reports from freelance journalists and independent investigators sharing unverified ballot chain-of-custody footage, while retaining officially sourced misinformation about voter fraud, because centralized trust mechanisms prioritized institutional sourcing over procedural accountability, revealing how the pursuit of informational stability systematically disables grassroots verification actors.

Evidentiary latency

In 2021, Facebook’s restrictions on posts discussing the origins of SARS-CoV-2 silenced early hypotheses about lab leaks—including those later corroborated by U.S. intelligence agencies—because its anti-misinformation protocols classified non-consensus scientific inquiry as harmful falsehood, demonstrating how real-time moderation systems cannot accommodate the slow, iterative timeline of investigative discovery.

Attribution asymmetry

When Twitter in 2022 enforced new rules against 'manipulated media' to block the New York Post’s Hunter Biden laptop story, it applied forensic verification standards to journalistic artifacts that were not required of state actors disseminating disinformation, exposing how platform policies designed to reduce harm privilege the provenance of the speaker over the veracity of the content, thus weaponizing authenticity standards against adversarial journalism.

Editorial Liability Shield

Platforms that retroactively penalize news outlets for factual reporting under misinformation policies distort liability norms, as seen when Meta downgraded and defunded outlets like The Local in Germany after they published leaked documents on migration statistics that contradicted official narratives. The mechanism is not content moderation per se but the imposition of reputational and economic sanctions on organizations whose sourcing meets journalistic standards, thus creating a chilling effect on legally protected speech. The non-obvious consequence is that platforms become de facto arbiters of state truth, undermining the press’s legally recognized role as a counterweight to institutional power.

Temporal Monopoly on Truth

Investigative journalism becomes misclassified as misinformation when platforms enforce real-time alignment with emergent consensus, as occurred during the 2021 U.S. Capitol breach when Twitter suspended journalists sharing unverified footage of protester demographics before official reports existed. The dynamic hinges on platforms privileging temporal proximity to institutional validation over evidentiary rigor, thereby criminalizing the first draft of history. This reframes censorship not as overt suppression but as the monopolization of truth-timing—where delay is weaponized against the press’s core function to precede official narratives.

Deeper Analysis

Where are the news outlets and journalists most affected by these moderation systems located, and how does that compare to where misinformation enforcement is strongest?

Digital Periphery

News outlets in the Global South—such as Kenya’s Nation Media Group or Indonesia’s Kompas—are disproportionately disrupted by automated content moderation systems designed in Silicon Valley, not because they produce more misinformation but because their linguistic and cultural contexts are poorly modeled by AI classifiers trained on Western datasets. These systems conflate regional discourse markers with disinformation, leading to erroneous demonetization and suppression, while enforcement remains most aggressive in high-income democracies like Germany or Canada, where platforms prioritize regulatory compliance. This reveals that moderation ‘strength’ maps not to misinformation prevalence but to geopolitical risk exposure for platforms, exposing a structural bias where harm is externalized onto linguistically diverse, less legally adversarial markets.

Explore further:

How often do platform algorithms block or delay investigative stories before they’re published, just because they look risky?

Shadow Moderation Thresholds

Platforms systematically delay investigative content from high-credibility journalism outlets by embedding opaque escalation rules in automated moderation layers. During the 2020 U.S. election, Facebook’s internal ‘Dangerous Individuals and Organizations’ (DIO) policy retroactively flagged investigative reports by ProPublica on militia networks as ‘potentially harmful,’ reducing their distribution by 70% within the algorithmic feed without notification. This mechanism operates through pre-publication content scoring systems that treat sourcing from known investigative outlets as a risk multiplier, revealing that credibility can inadvertently trigger suppression when conflated with radicalization signals in the algorithm’s risk calculus.

Preemptive Visibility Throttling

YouTube’s recommendation algorithm downgraded independent investigative videos about vaccine safety following algorithmic updates in 2019, even when published by credentialed medical journalists like those at The Greyzone. Internal documents leaked in 2021 showed that content matching behavioral clusters associated with ‘borderline misinformation’—regardless of factual accuracy—was automatically restricted in visibility, especially if it featured extended interviews or cited non-institutional experts. This demonstrates that algorithmic risk models penalize structural formats common in investigative journalism, such as deep sourcing or counter-narrative framing, exposing a systemic bias against methodological rigor when encoded as engagement patterns.

Crisis-State Content Purging

Twitter’s algorithmic suppression of investigative threads on Pegasus spyware in July 2021, particularly those by Amnesty International and Forbidden Stories, occurred not through direct censorship but via delayed propagation during peak relevance. Analysis by the Stanford Internet Observatory found that tweets containing forensic links to spyware code were deprioritized in real-time ranking due to sudden activity surges misclassified as platform manipulation. This reveals that crisis-responsive algorithmic modes interpret coordinated dissemination of investigative findings as synthetic behavior, effectively treating virality of truth-seeking as a system-level threat.

Algorithmic preemption

Investigative stories from WikiLeaks after 2016 were systematically throttled by Facebook’s newsfeed algorithms before publication visibility could build, not due to content removal but through ranking suppression rooted in updated ‘authenticity’ criteria that prioritized personal over institutional scrutiny. This shift marked a transition from reactive takedown regimes to anticipatory algorithmic dampening, wherein platforms began treating investigative outlets as engagement liabilities rather than permissible risks, especially following user fatigue and advertiser backlash after the 2016 U.S. election cycle. The non-obvious insight is that the blockage is not a denial-of-service but a degradation-of-reach, emerging not in censorship per se but in the recalibration of distribution incentives that deprioritized adversarial journalism as platform liability.

Explore further:

How did the way platforms handled election content in 2020 set a pattern for which voices get heard or silenced in later political moments?

Preemptive Moderation

Platforms' deployment of automated content filters before the 2020 election established preemptive moderation as the default response to political speech. Major social media companies like Facebook and Twitter used AI systems to flag or suppress posts containing disputed claims about voting procedures, ballot integrity, and candidate eligibility before human review, relying on advisories from organizations like the Cybersecurity and Infrastructure Security Agency (CISA). This mechanism normalized the delegation of political speech judgments to backend technical systems calibrated not to legal standards but to institutional risk thresholds, making suppression invisible and non-negotiable—what most users associate with 'censorship' was reframed as routine safety enforcement, obscuring the editorial power now embedded in algorithmic anticipation.

Crisis Credentialing

The coordinated labeling of election misinformation by platforms in real-time during vote counting established crisis credentialing as a new gatekeeping function in public discourse. When Twitter and Facebook applied warning tags to Donald Trump’s tweets and restricted sharing of specific articles, they positioned themselves not just as hosts but as arbiters of factual legitimacy during high-stakes political moments, using internal teams like Facebook’s Cross Check program and external fact-checkers vetted by the International Fact-Checking Network. This mirrored traditional media’s editorial role, yet operated at platform scale and speed—what users recognize as 'fact-checking' became a crisis tool that amplified institutional voices while sidelining those deemed unreliable by consensus bodies, entrenching a credential-based hierarchy in digital speech.

Where else are platforms pulling back on enforcement not because misinformation is lower but because the political or legal risks are smaller?

Regulatory shadow gradients

Platforms reduce enforcement in regions where national regulators lack transnational enforcement reciprocity, enabling content moderation retreats without legal exposure. This occurs because domestic agencies in countries like Indonesia or Nigeria may issue demands, but cannot compel compliance from Silicon Valley HQs absent bilateral enforcement treaties—creating a gradient of accountability shadows where platforms exploit jurisdictional asymmetries. The mechanism operates through asymmetric legal capacity in cross-border digital governance, a dynamic overlooked because most analyses focus on state sovereignty rather than the infrastructural gaps in enforcement ecosystems. Regulatory shadow gradients reveal that pullbacks are less about politics per se and more about the operational invisibility of weak regulatory networks in global compliance architectures.

Infrastructure opacity leverage

Platforms ease enforcement in markets where internet backbone dependencies are state-controlled and monitoring capabilities are obscured, such as in Ethiopia or Kazakhstan, because operators cannot verify content provenance when national ISPs filter or throttle traffic upstream. This retreat happens not due to reduced misinformation, but because the platform’s detection systems become blind when routing passes through opaque sovereign networks that compress or scramble metadata. The phenomenon is sustained by the hidden dependency of AI moderation on clean network telemetry, which most assessments ignore in favor of speech policy debates. Infrastructure opacity leverage thus reframes enforcement pullbacks as technical surrender, not political calculation.

Moderator labor geography

Content enforcement wanes in linguistic zones where qualified local moderator labor is scarce and offshoring is politically sensitive, such as in Indigenous or minority language contexts in Latin America, because platforms cannot scale human review without importing foreign contractors who lack cultural fluency. The pullback stems not from misinformation decline but from an unmet dependency on geographically concentrated expertise that unions or national laws restrict, as seen in Bolivia’s restrictions on foreign digital labor. Most analyses miss that enforcement intensity is bottlenecked by the political economy of outsourced moderation labor, not algorithmic capacity. Moderator labor geography exposes enforcement as a supply chain function, not just a policy dial.

Compliance Fatigue

Platforms reduced enforcement in Eastern Europe after 2022 not because misinformation declined but because geopolitical alignment with NATO states reduced regulatory retaliation risk. As the conflict in Ukraine normalized wartime information controls, platforms faced less penalty for inaction when allies framed content moderation as secondary to intelligence and cyberwarfare coordination. This shift from 2016–2020, when platforms preemptively scrambled to avoid U.S. congressional backlash, reveals how crisis-driven legitimization of state narratives enabled platforms to outsource accountability. The enduring artifact is a tolerance for enforcement drift when government priorities redefine harm.

Moderation Arbitrage

After 2018, platforms scaled back Arabic-language misinformation teams while maintaining English ones, not because disinformation volume dropped but because U.S. regulatory focus receded from MENA regions. When the post-2016 election panic subsided, legal incentives narrowed to domestic politics, letting platforms reallocate moderation away from Yemen or Syria where enforcement carried higher operational risk and negligible domestic scrutiny. This strategic de-emphasis—unlike prior universalist scaling pledges—exposes how platforms redefined 'compliance' as geographically tiered. The result is a sustained hierarchy where enforcement follows legal exposure, not harm density.

Accountability Evasion

Platforms withdrew algorithmic transparency measures in India after 2021 after local courts stopped treating misinformation as a justiciable harm, shifting from pre-2019 norms when judicial activism pressured moderation. As Indian judiciary retreated from intervening in viral lynchings linked to viral content, platforms recalibrated enforcement to match the lowered institutional expectation, not improved information quality. Unlike in Brazil or Germany, where courts enforced takedowns, India’s judicial abstention became a signal to minimize investment. This retreat produced a regime where platforms measure risk not by virality of falsehoods but by judicial enforceability, not civic consequence.

Regulatory Risk Discount

Indian social media platforms relaxed misinformation enforcement after the 2019 general election amidst reduced risk of government reprisal for non-compliance, despite persistent viral falsehoods; this shift was not due to improved information quality but to the calculus that stricter enforcement during electoral sensitivity carried higher political cost than post-election tolerance, revealing how compliance urgency is dynamically priced against state pressure rather than harm thresholds.

Content Moderation Arbitrage

Facebook decreased enforcement intensity in the Philippines under the Duterte administration’s tenure, maintaining minimal action against coordinated inauthentic behavior linked to state actors not because misinformation declined but because corporate risk assessment downgraded legal exposure after local regulatory threats were temporarily defused through negotiated access agreements, exposing how platform enforcement effort is reallocated across jurisdictions like a risk-hedging portfolio.

Judicial Exit Incentive

YouTube reduced removals of extremist content in Germany after 2020 not because misinformation waned but because the enforcement burden shifted from platform discretion to state-appointed oversight bodies under the NetzDG law, allowing the company to scale back proactive detection while remaining compliant, thus revealing that platforms retreat when legal accountability can be institutionally outsourced to absorb political liability.

Explore further:

Who decides which voices count as credible when platforms use behavior patterns to flag misinformation, and how does that affect journalists challenging official narratives?

Epistemic arbitrage

Platform algorithms designate credibility through behavioral proxies like engagement velocity and network clustering, privileging institutional sources that generate stable, consensus-aligned patterns—thus, actors such as CDC or BBC are algorithmically amplified while outlier patterns from investigative journalists are demoted, not due to factual accuracy but structural alignment with dominant epistemic networks. This occurs because automated trust signaling in platforms depends on recurrence and scale of validation, which state-affiliated or mainstream entities can reliably produce, whereas adversarial journalism often emerges asynchronously and from sparse networks. The non-obvious consequence is that credibility becomes a function of temporal and networked predictability, not truth verification, enabling platforms to outsource editorial judgment to invisible statistical regularities that systematically disadvantage challenges to official narratives.

Infrastructural capture

The design standards of content moderation infrastructures—such as Facebook’s Third-Party Fact-Checking Program and Twitter’s Community Notes—are shaped by partnerships with established media and governmental risk frameworks, making credibility decisions contingent on prior institutional recognition rather than argumentative validity. Because these systems rely on pre-approved organizations to label content, journalists working outside consortia like the International Fact-Checking Network (IFCN) lack procedural access to contest misclassification, embedding a regulatory barrier within technical architecture. The underappreciated dynamic is that technical neutrality is compromised upstream by credentialing regimes that equate legitimacy with organizational pedigree, effectively allowing legacy institutions to colonize algorithmic governance and insulate official narratives from real-time journalistic scrutiny.

Narrative debt

When platforms use historical behavior patterns to model misinformation risk, they encode past speech environments into present credibility assessments, disproportionately penalizing journalists who specialize in exposing concealed power structures because their prior work often triggered coordinated attacks or mass reporting by state-linked actors. This creates a feedback loop where suppression events are misread as evidence of low credibility rather than political targeting, as seen in the downranking of independent Ukrainian or Belarusian reporters during periods of hybrid warfare. The overlooked mechanism is that behavioral models conflate resilience under attack with inherent falsity, treating high-reporting thresholds as intrinsic to the source rather than recognizing them as external aggression—thus rewarding conformity and punishing investigative tenacity.

Platform Epistemic Authority

Social media platforms in the Global North, particularly U.S.-based ones like Meta and X, decide which voices count as credible by embedding Western liberal democratic norms into algorithmic behavior patterns that prioritize verified accounts, institutional sources, and consistency with established fact-checking networks. This mechanism privileges government health agencies, mainstream news outlets, and credential-based expertise—especially during crises like pandemics or elections—while marginalizing alternative narratives, even when those are widely accepted in non-Western contexts. The non-obvious consequence, under the lens of Familiar Territory, is that what feels like neutral technical moderation—such as flagging ‘inauthentic behavior’—actually enforces a culturally specific epistemology where trust is institutionalized rather than relational or communal, disrupting journalists in the Global South who draw legitimacy from local consensus or spiritual authority rather than peer review.

Narrative Sovereignty

Governments in regions like China, Iran, and Russia define credibility by aligning behavior-based misinformation flags with state-determined truths, treating journalism as a function of national loyalty rather than adversarial inquiry. Platforms operating within these jurisdictions either comply with or are replaced by domestic alternatives—like WeChat or Sputnik—that measure behavioral patterns such as collective sentiment shifts or loyalty signaling, not factual accuracy in the Western sense. The overlooked dynamic here, despite public familiarity with censorship, is that these systems do not merely suppress dissent but actively promote an alternate conception of truth maintenance, where journalists challenging official narratives aren’t excluded for spreading falsehoods but for violating cultural expectations of social harmony and political cohesion, recasting credibility as an act of allegiance.

Colonial Moderation Logic

Western content moderation frameworks, when applied globally, treat non-Western belief systems—such as Indigenous oral traditions or Islamic scholarly consensus—as behavioral anomalies because they often circulate through decentralized, kinship-based networks that lack timestamped citations or institutional branding. This leads platforms to flag content from journalists in places like Nigeria or Indonesia as suspicious when they amplify narratives rooted in communal memory or religious interpretation, even if locally authoritative. What remains unseen, despite widespread recognition of bias in tech, is that the very pattern of ‘credibility signaling’—like repeated engagement from verified users—is itself a Western social ritual, making the global enforcement of such metrics a form of epistemic imperialism where the familiar becomes the universal, and alternative modes of truth-telling are algorithmically erased not because they are false, but because they behave differently.

Infrastructural Authority

The credibility of voices in misinformation systems is ultimately decided by the private engineering teams at major tech platforms who operationalize behavior patterns into automated enforcement rules, because they embed normative judgments about trustworthiness into algorithmic thresholds—such as what constitutes 'coordinated inauthentic behavior'—without public oversight, making their design choices function as de facto speech governance. This shift hides power within technical execution, where decisions about whose journalism appears manipulative are made not by editors or fact-checkers but by backend risk assessments calibrated to avoid geopolitical backlash, rendering adversarial reporting vulnerable if it mimics the temporal or network signatures of disinformation. The non-obvious reality is that credibility is now policed through infrastructure, not editorial review, privileging stability over dissent.

Narrative Immunization

Credibility is effectively allocated to voices that align with existing institutional risk appetites, meaning journalists challenging official narratives are systematically disadvantaged because their behavior—such as rapid source triangulation or reliance on leaked data—triggers pattern-based flags designed to detect 'malicious outliers.' Platforms rely on baseline behavioral norms derived from historically compliant media, so investigative work that disrupts those norms is misclassified as synthetic or manipulative, even when accurate, reinforcing a feedback loop where only non-threatening dissent survives moderation filters. This reveals that the defense against misinformation has become a mechanism for narrative immunization, protecting power not through censorship but through behavioral profiling.

Epistemic Arbitrage

The determination of credible voices is increasingly outsourced to third-party verification ecosystems—such as cross-platform data pools and 'trusted' government advisories—whose participation dictates whether a journalist’s behavior is interpreted as legitimate or suspect, privileging those embedded in Western-aligned information networks. Journalists operating outside these circuits, especially in conflict zones or authoritarian contexts, face higher scrutiny because their behavioral patterns—like sudden spikes in engagement or use of anonymous sources—lack the validating metadata that platforms use to distinguish truth from manipulation, effectively turning access to infrastructure into epistemic capital. This exposes how credibility has become subject to epistemic arbitrage, where truth attribution follows logistical affiliation, not evidentiary merit.

Relationship Highlight

Infrastructural Authorityvia Clashing Views

“The credibility of voices in misinformation systems is ultimately decided by the private engineering teams at major tech platforms who operationalize behavior patterns into automated enforcement rules, because they embed normative judgments about trustworthiness into algorithmic thresholds—such as what constitutes 'coordinated inauthentic behavior'—without public oversight, making their design choices function as de facto speech governance. This shift hides power within technical execution, where decisions about whose journalism appears manipulative are made not by editors or fact-checkers but by backend risk assessments calibrated to avoid geopolitical backlash, rendering adversarial reporting vulnerable if it mimics the temporal or network signatures of disinformation. The non-obvious reality is that credibility is now policed through infrastructure, not editorial review, privileging stability over dissent.”

Similar Queries in Other Domains

Analysis: Explore the causal links and hidden assumptions behind forcing platforms to tackle lies or boost competing media — map the reasoning chains interactively.
Should We Force Platforms to Tackle Lies or Boost Competing Media?
Explore the causal links and hidden assumptions behind forcing platforms to tackle lies or boost competing media — map the reasoning chains interactively.
Analysis: Explore the journalists dilemma between quick tweets and fact-checked news — trace the trade-offs and unpack hidden biases interactively.
Quick Tweets or Fact-Checked News? The Journalists Dilemma
Explore the journalists dilemma between quick tweets and fact-checked news — trace the trade-offs and unpack hidden biases interactively.
Analysis: Explore the real-time news vs misinformation landscape on Twitter for journalists — trace causal links and unpack hidden biases interactively.
Real-Time News vs Misinformation on Twitter for Journalists
Explore the real-time news vs misinformation landscape on Twitter for journalists — trace causal links and unpack hidden biases interactively.
Analysis: Explore how economic data credibility differs in editorials versus startup dashboards — unpack assumptions and trace causal links interactively.
Is Economic Data More Credible in Editorials or Startup Dashboards?
Explore how economic data credibility differs in editorials versus startup dashboards — unpack assumptions and trace causal links interactively.

Similar Concepts in Other Domains

Analysis: Explore the vaccine safety debate by tracing media literacy impacts and unpacking platform design effects — interactively map causal links and assumptions.
Vaccine Safety Debate: Media Literacy or Platform Design?
Explore the vaccine safety debate by tracing media literacy impacts and unpacking platform design effects — interactively map causal links and assumptions.
Analysis: Explore the complex web of Reddits community moderation — trace causal links and unpack hidden assumptions to assess its true impact on epistemic damage.
Does Reddits Community Moderation Truly Fight Epistemic Damage?
Explore the complex web of Reddits community moderation — trace causal links and unpack hidden assumptions to assess its true impact on epistemic damage.
Analysis: Explore the asymmetry of power in content moderation — unpack the defenses for relying on platforms’ transparency reports through interactive 3D mapping.
Is Platform Transparency Enough to Balance Power in Content Moderation?
Explore the asymmetry of power in content moderation — unpack the defenses for relying on platforms’ transparency reports through interactive 3D mapping.