Digital-Era Psychological Operations and the Information Environment

Full report

On this page

Introduction and Conceptual Framework

The digitalization of the global information environment has fundamentally transformed the mechanisms, scale, and velocity of psychological operations (PSYOPS), public influence campaigns, and propaganda. The intersection of networked platforms, algorithmic distribution, data-driven advertising, cyber-enabled information operations, and generative artificial intelligence (AI) has created an ecosystem where both state and non-state actors can manipulate public perception with unprecedented precision. The contemporary threat landscape is no longer defined merely by the broadcast of false information, but by the systemic exploitation of cognitive vulnerabilities through computational architecture. This report provides an exhaustive analysis of the digital-era information environment, separating the technical infrastructure of influence operations from their cognitive effects. It explores the historical evolution of computational propaganda, the telemetry challenges inherent in platform analysis, the paradigm-shifting impact of synthetic media, and the localized defensive postures required to build cognitive resilience in communities such as Cicero, Illinois. To establish evidentiary precision and analytical rigor, the lexicon of digital influence must be rigorously defined and distinguished. Conflating these terms routinely leads to analytical failures and flawed policy responses. Foreign Information Manipulation and Interference (FIMI) refers to intentional, coordinated activities executed by state or state-linked actors that lie at the intersection of national security, information integrity, and hybrid warfare1. FIMI encompasses a broad spectrum of manipulative behaviors aimed at distorting discourse, conceptually overlapping with but distinct from foreign influence, which may include overt and legally permissible diplomatic or broadcasting efforts. Coordinated Inauthentic Behavior (CIB) is a term popularized by platform security teams to describe networks of accounts and pages working together to spread specific messages while concealing their true nature and origins2. CIB focuses strictly on the behavior and coordination of the network rather than the underlying truth, falsity, or geopolitical origin of the content. The operational execution of these campaigns relies on computational propaganda, which is the use of algorithms, automation, and human curation to purposefully distribute misleading information over social media networks. Within this domain, a distinction must be made between bots and troll farms. A bot is an automated or semi-automated script designed to execute programmatic actions—such as liking, sharing, or posting—at scale, often to manipulate trending algorithms. Conversely, a troll farm relies on human operators, frequently working in coordinated shifts, to manually inject inflammatory, deceptive, or hyper-partisan content into digital discourse. These human operators provide a level of cultural nuance and contextual awareness that traditional bots historically lacked3. Both entities frequently engage in astroturfing, the practice of masking the sponsors of a message or organization to make it appear as though it originates from and is supported by organic, grassroots participants. Furthermore, influence operations weaponize the financial and architectural mechanisms of digital platforms through dark advertising and microtargeting. Dark advertising refers to hyper-targeted digital advertisements visible only to the publisher and the specific micro-targeted audience, circumventing public scrutiny and standard counter-speech mechanisms5. Microtargeting utilizes consumer data and psychographics to identify the specific vulnerabilities or interests of individuals or small groups, influencing their thoughts through precisely tailored messaging. When users attempt to verify this information, they frequently encounter search manipulation, wherein operators exploit search engine algorithms to ensure that specific, conspiratorial, or state-backed narratives dominate the top search results, a tactic particularly effective when exploiting "data voids" surrounding emerging news events. The deployment of these tactics is increasingly augmented by deepfakes and synthetic media—highly realistic, AI-generated audio, video, or image content designed to depict events that never occurred or statements that were never made. To combat the proliferation of synthetic media, the technology and publishing industries rely on provenance systems. Provenance refers to the verifiable history of a digital asset's origins and alterations. The Coalition for Content Provenance and Authenticity (C2PA) has established an open technical standard that embeds cryptographically signed metadata into media files to track their creation and edit history, offering a mechanism to verify digital assets at the point of consumption6.

The Analytical Separation Model

To accurately attribute and analyze influence operations without succumbing to hyperbole or analytical conflation, the digital ecosystem must be systematically divided into eight distinct components. Failing to separate these elements leads to flawed intelligence, misattribution, and the implementation of ineffective countermeasures.

The rigid separation of Reach and Demonstrated Effect is the most critical conceptual boundary in evaluating digital operations. High reach—often manifesting as millions of impressions—frequently correlates with zero demonstrated cognitive or behavioral effect. Automated accounts heavily amplify content, but often this content enters isolated bot-echo chambers or is entirely ignored by organic users12. Assuming that impressions equal persuasion is a fundamental flaw that overestimates the efficacy of foreign interference and downplays the resilience of domestic populations.

System History: The Evolution of the Information Environment

The contemporary digital information environment is the result of a multifaceted evolution in cyber capabilities, platform architecture, and psychological warfare doctrine spanning over two decades. Tracing this history reveals a continuous arms race between offensive information operations and platform defense mechanisms. During the era of early forums and search manipulation (2000–2010), digital influence operations focused on traditional web infrastructure. State and non-state actors utilized message boards, blog comments, and basic search engine optimization (SEO) techniques to manipulate public perception. The primary objective was to dominate early search engine results by exploiting data voids—niche search terms lacking authoritative content—and filling them with heavily optimized, partisan, or state-sponsored narratives. The landscape shifted dramatically with the rise of social bots and microtargeting (2010–2016). The mass adoption of social media networks transitioned the paradigm toward algorithmic exploitation. Automated scripts were deployed to hijack trending topics and simulate artificial consensus. Concurrently, the maturation of data-driven advertising allowed for granular psychographic microtargeting. Operations during this era weaponized the inherent openness of platforms, injecting inflammatory material directly into specific demographic feeds to exploit political, religious, and societal divides16. The period of hack-and-leak operations and state media networks (2014–2018) integrated traditional cyber espionage with digital propaganda. Operations such as the targeting of the Democratic National Committee (DNC) in the United States or the Emmanuel Macron campaign in France relied on hacking private communications and laundering them through decentralized digital actors. State media networks, such as RT and Sputnik, provided a veneer of journalistic legitimacy by amplifying these stolen materials to global audiences17. This blending of kinetic cyber effects with psychological operations was also evident in Ukraine, where cyberattacks on energy infrastructure were synchronized with disinformation campaigns to maximize civilian panic19. As major social media platforms invested heavily in detecting CIB, operators adapted through encrypted messaging and ephemeral tactics (2018–2021). Influence networks shifted toward closed, encrypted messaging applications where researchers and platform telemetry could not penetrate. Concurrently, sophisticated campaigns utilized "ephemeral disinformation." Operators would post spoofed articles, artificially amplify them to gain media attention, and quickly delete the content, redirecting the URL to a legitimate news site20. This technique created an appearance of legitimacy while intentionally destroying the forensic evidence required for attribution. The current epoch is defined by generative AI and platform decentralization (2022–Present). The integration of Large Language Models (LLMs) and diffusion models has fundamentally altered the economics of computational propaganda. Generative AI allows for the creation of culturally nuanced, linguistically flawless text and highly realistic synthetic media at scale, overcoming the traditional limits of manual troll farms21. Simultaneously, the fracturing of the social media landscape—moving from centralized town squares to decentralized protocols and niche platforms—has fragmented audience attention and severely complicated unified defensive monitoring and threat intelligence sharing.

The Platform-Telemetry Problem

A central challenge in evaluating digital-era psychological operations is the "platform-telemetry problem." Analysts, policymakers, and the public rely heavily on incomplete, asymmetric data to make sweeping assessments about the information environment, often leading to skewed perceptions of threat and influence. The foundation of this problem is proprietary evidence and access asymmetry. The vast majority of forensic data required to definitively attribute an influence operation—such as IP addresses, device identifiers, creation timestamps, internal coordination patterns, and private messaging metadata—is held exclusively by technology platforms12. Independent researchers, academics, and civil society organizations are largely confined to analyzing public-facing Application Programming Interfaces (APIs), scrapeable surface data, or curated datasets provided voluntarily by the platforms after an operation has been disrupted. This access asymmetry means that public understanding of computational propaganda is heavily reliant on corporate transparency, which frequently fluctuates based on business priorities, legal liabilities, and public relations strategies. This asymmetry contributes directly to the disconnect between views and belief change. Platform telemetry fundamentally measures engagement—clicks, likes, retweets, and impressions—not cognitive impact or persuasion. A persistent analytical flaw in media reporting is equating millions of bot-generated impressions with widespread psychological manipulation. Automated accounts frequently interact primarily with each other in isolated networks. For instance, in the Pentagon's Unheard Voice campaign, researchers noted that the vast majority of posts received almost no organic engagement; the fake accounts were effectively talking to themselves, generating telemetry without influence12. The sheer volume of synthetic content does not intrinsically translate to human behavioral shifts. Compounding these issues is the inconsistency in labels and the attribution trap. Efforts to mathematically track influence operations frequently encounter severe false-positive loops, leading to the misattribution of organic domestic polarization to foreign state actors. The most prominent historical example is the Hamilton 68 dashboard. Created by the Alliance for Securing Democracy, the dashboard claimed to track "Russian-linked" influence networks on Twitter and was highly influential in shaping media narratives about foreign interference16. However, subsequent disclosures (the "Twitter Files") revealed that the proprietary list of accounts tracked by the dashboard largely consisted of domestic American citizens engaged in organic, albeit highly partisan, political discourse25. This telemetry failure illustrates how the anxiety surrounding foreign influence can be weaponized, delegitimizing authentic domestic speech through black-box attribution models that lack transparent, peer-reviewed methodologies.

AI and Synthetic Media: The Automation of Influence

The advent of generative AI represents a phase transition in computational propaganda. Historically, influence operations faced a strategic trilemma: they could achieve massive scale, linguistic and cultural authenticity, or low operational cost—but rarely all three simultaneously. Generative AI fundamentally solves this trilemma, allowing operators to produce high-fidelity, customized propaganda at negligible costs and unprecedented volume. Generative AI provides unparalleled capabilities in scale, persona consistency, and content generation. Previous state-backed operations required physical infrastructure filled with salaried employees manually translating, drafting, and scheduling posts. Today, AI frameworks can generate thousands of unique, contextually appropriate responses per minute. Furthermore, AI allows for strict persona consistency. An LLM can be prompted to maintain a specific synthetic identity—complete with localized slang, consistent ideological leanings, and appropriate emotional responses—across thousands of interactions without suffering the fatigue or human error that previously exposed manual troll farm operators21. The 2024 disruption of the Russian Meliorator bot farm demonstrated this exact capability, utilizing an AI-enhanced software package to create authentic-appearing personas en masse, which automatically aligned their posts to mimic ordinary users and interact directly with target audiences11. To defend against the proliferation of synthetic media, the technology industry has coalesced around provenance systems, most notably the Coalition for Content Provenance and Authenticity (C2PA). C2PA operates on traditional Public Key Infrastructure (PKI)—similar to X.509 certificates used in secure web browsing—to embed cryptographically signed, tamper-evident metadata manifests into media files6. When a digital asset is created, whether via a hardware-secured camera or an AI generator, the C2PA manifest records the origin, author, and any subsequent edits6. Because the manifest is cryptographically bound to the content, any unauthorized tampering breaks the digital signature, alerting the viewer that the content is unverified6. Importantly, C2PA does not rely on blockchain technology, allowing for faster, offline verification without transaction fees6. Despite its cryptographic rigor, C2PA contains inherent vulnerabilities regarding metadata loss. The standard is designed for transparency, not Digital Rights Management (DRM). Consequently, C2PA metadata can be intentionally or unintentionally stripped via basic digital actions, such as taking a screenshot, uploading the file to a platform that compresses and re-encodes media, or stripping EXIF data to protect user privacy6. While the absence of C2PA data on a political video can serve as a warning sign, it cannot definitively prove whether the underlying media is real or synthetic. To counter metadata stripping, organizations like OpenAI supplement C2PA with SynthID, an invisible watermarking technology embedded directly into the pixels or audio waves, which can survive some degree of compression or cropping28. Furthermore, provenance systems struggle with the nuance of human editing, leading to severe false positives. In early 2024, Meta implemented an automated system to tag content containing AI metadata with a "Made with AI" label30. This triggered massive backlash from professional photojournalists and digital artists, whose authentic, real-world photographs were flagged as "Made with AI" simply because they used standard software tools featuring minor AI enhancements, such as Adobe Photoshop's generative fill or basic retouching32. Acknowledging that the binary label failed to capture the spectrum of digital editing and unnecessarily damaged the credibility of authentic creators, Meta altered the label to "AI info" on July 1, 2024, allowing users to click for deeper context on how AI was utilized in the workflow32. This incident underscores the difficulty of deploying automated AI-assisted moderation at scale without generating false positives that erode user trust37.

Case Studies in Digital Influence

A robust understanding of the digital information environment requires an examination of specific, well-documented operations. The following twelve case studies demonstrate the spectrum of tactics, actors, and outcomes in modern computational propaganda, highlighting the origins, evidence, reach, effect, and correction history of each campaign.

Table 1: Case Studies of Digital-Era Influence Operations

1\. Internet Research Agency (IRA): Sponsored by Russian oligarch Yevgeny Prigozhin, who openly admitted to the operation in 2022, the IRA utilized human troll farms, persona hijacking, and sophisticated social media microtargeting to inject inflammatory racial and political content into the US information space4. The operation achieved a highly demonstrated effect by polarizing discourse, successfully organizing offline protests, and manipulating mainstream media into amplifying their constructed narratives. 2\. Doppelganger (SDA/Structura): Operating under the direction of the Russian Presidential Administration via firms like the Social Design Agency (SDA) and Structura, Doppelganger created cloned versions of legitimate news websites (e.g., creating spiegel.ltd to mimic Der Spiegel)5. The technical infrastructure relied heavily on European "bulletproof" hosting providers like Aeza to maintain resilience8. While the campaign saturated platforms with pro-Russian narratives, reaching millions of users, its actual cognitive effect is debated, though it resulted in significant Western sanctions and infrastructure seizures5. 3\. Spamouflage (Dragonbridge): This massive, state-aligned Chinese operation relies on vast networks of disposable accounts to spam pro-Beijing narratives across global platforms41. Spamouflage relies heavily on algorithmic manipulation through sheer volume. However, because the content often lacks cultural nuance and the bot behavior is rudimentary, platform telemetry frequently detects and removes the networks, resulting in enormous generated volume but historically negligible organic reach or cognitive influence. 4\. Meliorator: In July 2024, the US Department of Justice disrupted an AI-enhanced bot farm operated by affiliates of RT and Russian intelligence (FSB)11. Utilizing a software package called Meliorator, the operators created fictitious personas en masse, utilizing open-source tools like Faker for profile images and web crawlers to build realistic biographies11. The software enabled bots to seamlessly mimic ordinary users and interact directly with political candidates. The disruption involved seizing the underlying domains (MLRTR.com and OTANMAIL.com), severely degrading the operation's technical infrastructure27. 5\. US Military Anti-Vax Campaign (ChinaAngVirus): Between 2020 and 2021, the US Department of Defense operated a covert disinformation campaign targeting the Philippines and Central Asia to counter China's COVID-19 diplomacy43. Using covert personas, the campaign utilized hashtags like \#ChinaIsTheVirus to cast doubt on the efficacy of the Sinovac vaccine and Chinese medical aid43. Exposed by Reuters in June 2024, the operation demonstrated high reach and severe effects, heavily contributing to vaccine hesitancy in the region. The campaign was eventually terminated following an internal review by the Biden administration43. 6\. Unheard Voice: In 2022, researchers from Graphika and the Stanford Internet Observatory exposed a massive pro-Western influence operation operating in the Middle East and Central Asia, linked to US Central Command (CENTCOM)15. The operation utilized GAN-generated faces and spoofed news outlets to push anti-Russia and anti-China narratives12. However, platform telemetry revealed that despite running for five years, the campaign largely failed to build organic audiences, with covert assets receiving minimal engagement, highlighting the limitations of inauthentic behavior without compelling content12. 7\. Endless Mayfly: An Iran-aligned disinformation campaign active since 2016, Endless Mayfly specialized in "ephemeral disinformation." The operators would spoof domains of legitimate institutions (such as the Belfer Center or Le Soir), publish fabricated articles, and use inauthentic personas to amplify the links to journalists and dissidents20. Once the narrative gained traction, the operators deleted the fake articles and redirected the URLs to the legitimate sites being impersonated. This technique effectively laundered the source of the disinformation, inducing lasting media errors and confusion even after the technical evidence was destroyed14. 8\. Slovakian Election Deepfake: Days before the highly contested 2023 Slovakian parliamentary elections, a generative audio deepfake circulated on platforms like Facebook and Telegram. The audio depicted Michal Simecka, leader of the Progressive Slovakia party, allegedly discussing a scheme to buy votes from the Roma minority49. Although AFP fact-checkers quickly identified the audio as synthetic, it was deployed during a 48-hour media blackout period preceding the vote, severely limiting the ability of traditional media to correct the record49. This demonstrated the devastating effect of timing in synthetic media deployment. 9\. Hamilton 68 Dashboard: Created by the Alliance for Securing Democracy, the Hamilton 68 dashboard purported to track 600 "Russian-linked" Twitter accounts to monitor foreign influence16. For years, the dashboard was cited by major media outlets to claim that Russian bots were driving domestic political controversies. However, internal communications released during the "Twitter Files" revealed that the proprietary list consisted predominantly of authentic, domestic American users expressing partisan viewpoints25. This case serves as a critical example of telemetry failure and the dangers of opaque, proprietary attribution mechanisms. 10\. Doppelganger Israel: In 2023, the Russian Doppelganger infrastructure pivoted its operations to target the Israeli electorate. The campaign injected political messaging designed to deepen existing domestic polarization within Israel, aiming to influence decision-makers to withhold military aid to Ukraine and distance Jerusalem from the United States13. While the operation operated at a low intensity, it demonstrated the agility of established FIMI infrastructure to rapidly re-target new geopolitical crises using the same technical foundation13. 11\. Nunes Memo (\#ReleaseTheMemo): In early 2018, a massive social media campaign utilizing the hashtag \#ReleaseTheMemo emerged, demanding the declassification of a controversial congressional memo51. The Hamilton 68 project initially claimed that Russian-linked bots were responsible for spreading the hashtag, driving widespread media panic. However, internal telemetry from Twitter confirmed that there was no significant Russian activity connected to the hashtag; the phenomenon was entirely driven by authentic domestic partisans and automated domestic amplifiers51. 12\. DTEK Hack-and-Leak: In the lead-up to and duration of the Russian invasion of Ukraine, Russian state cyber actors consistently targeted Ukrainian critical infrastructure. In 2022, DTEK, Ukraine's largest private energy investor, suffered a massive cyberattack on its IT infrastructure19. This kinetic cyberattack was paired with coordinated information operations designed to broadcast the vulnerability of the Ukrainian state and erode civilian morale, demonstrating the modern convergence of cyber warfare, infrastructure degradation, and psychological operations19.

Defensive and Public-Interest Focus: A Localized Approach in Cicero, Illinois

The proliferation of algorithmic propaganda, FIMI, and synthetic media cannot be defeated strictly through centralized platform moderation or reactive government takedowns. Content moderation at scale suffers from insurmountable challenges, including context collapse, language barriers, and the inevitability of high-profile false positives when utilizing AI-assisted moderation frameworks37. Consequently, building localized cognitive resilience and media literacy is the most effective and scalable defensive posture. This strategic shift from centralized censorship to decentralized resilience is actively demonstrated at the municipal and educational levels, specifically within Cicero, Illinois.

Media Literacy and Legislative Mandates

In a pioneering legislative move, Illinois enacted Public Act 102-0055, mandating that all public high schools incorporate a unit of media literacy into their curriculum beginning in the 2022-23 school year55. The mandate requires instruction on evaluating media messages, analyzing digital environments, creating media responsibly, and understanding the social responsibility of media consumption55. At the local level, J. Sterling Morton High School District 201—which serves Cicero, Berwyn, Lyons, Stickney, and McCook—integrates this mandate directly into its English Language Arts, Media Services, and Career and Technical Education curricula57. By embedding digital citizenship into the academic framework, District 201 equips students with the tools necessary to navigate the digital-era PSYOPS environment. Course offerings such as "Digital Literacy" prepare students to use technology responsibly, explicitly teaching the legal and ethical dimensions of respecting creative work, the risks of oversharing, and the mechanics of the internet60. Furthermore, courses in "Broadcast Journalism" and "Contemporary Literature" foster critical thinking, requiring students to analyze and communicate information effectively within an ever-evolving global society57. The district reinforces these concepts through its Student Handbook, which explicitly aligns student-produced media with the media literacy curriculum mandate (105 ILCS 5/27-20.08), ensuring that students practice transparency, source attribution, and ethical journalism58.

Lateral Reading, Origin Tracing, and Privacy

A cornerstone of the defense against computational propaganda is the pedagogical shift toward "lateral reading." Developed by researchers at the Stanford History Education Group, lateral reading abandons the outdated model of evaluating a website's credibility based on its internal design, URL, or "About" page—tactics easily spoofed by sophisticated operations like Doppelganger or Endless Mayfly5. Instead, lateral reading requires users to immediately leave the source in question, open new browser tabs, and consult external, trusted authorities to verify the original site's credibility, funding, and reputation61. In a localized context like Cicero, empowering students and the broader public to utilize lateral reading, combined with an understanding of AI provenance tools (such as recognizing when C2PA metadata has been suspiciously stripped), shifts the defensive burden from reactive platform takedowns to proactive cognitive immunity. By prioritizing platform transparency, origin tracing, and privacy awareness, localized education structures build a citizenry capable of identifying synthetic media and computational propaganda without relying on the delayed or opaque interventions of technology platforms.

Site Deliverables and Structured Data

To operationalize the findings of this report, the following structural deliverables are designed for integration into organizational knowledge bases.

Directory Structure

• /psyops/digital-era: Primary repository for historical evolution, FIMI definitions, and the 8-part attribution framework mapping technical infrastructure to cognitive effect.
• /psyops/platforms: Repository detailing platform telemetry, access asymmetry, algorithmic distribution vectors, dark advertising, and AI-assisted moderation challenges (including false-positive management).

Digital Influence Attribution Workbench

The Attribution Workbench provides a standardized schema for analyzing suspected operations without relying exclusively on proprietary platform data.

Synthetic-Media Timeline

Conclusion

The digital information environment has evolved from the era of manual message-board manipulation into a highly automated, AI-driven landscape. Operations orchestrated by state actors, military intelligence apparatuses, and domestic partisans alike leverage the scale of algorithmic distribution and synthetic media to fracture consensus and influence geopolitical outcomes. However, the technical sophistication of an operation—whether it is the ephemeral domain spoofing of Endless Mayfly, the AI-generated personas of Meliorator, or the infrastructure-cloning of Doppelganger—does not guarantee cognitive success. The primary vulnerability of the current ecosystem is not merely the existence of computational propaganda, but the telemetry asymmetry that prevents objective measurement of its impact. Misattribution, false positives in AI labeling, and the conflation of automated impressions with genuine human belief change present significant risks to democratic discourse. Mitigating these threats requires a transition away from an exclusive reliance on centralized platform moderation and opaque attribution dashboards. Instead, establishing robust cryptographic provenance standards, acknowledging the limitations of metadata durability, and institutionalizing cognitive defense mechanisms—such as the media literacy and lateral reading curricula being pioneered in districts like Cicero, Illinois—will be paramount in securing the integrity of the future information environment.

Works cited

1. Foreign Information Manipulation and Interference (FIMI), https://www.disinformation.ch/EU\_Foreign\_Information\_Manipulation\_and\_Interference\_(FIMI).html

2. 4th EEAS Report on Foreign Information Manipulation and Interference Threats \- European Union, https://www.eeas.europa.eu/sites/default/files/2026/documents/EEAS%204th%20Threat%20Report\_web%20version\_1.pdf

3. The NATO Association of Canada Disinformation and Digital Democracies in the 21st Century, https://natoassociation.ca/wp-content/uploads/2019/10/NATO-publication-.pdf

4. Russian oligarch Yevgeny Prigozhin, "Putin's chef," admits interference in U.S. elections, https://www.cbsnews.com/news/russia-us-election-interference-yevgeny-prigozhin-putin-chef-oligarch/

5. To be continued: The pro-Kremlin disinformation campaign Doppelganger in Germany \- CeMAS, https://cemas.io/en/publications/to-be-continued-doppelganger-germany/241115\_cemas\_report\_06\_en\_digital.en.pdf

6. Frequently Asked Questions (FAQ) \- C2PA, https://c2pa.wiki/getting-started/faq/

7. C2PA Technical Specification, https://spec.c2pa.org/specifications/specifications/1.0/specs/C2PA\_Specification.html

8. Russian disinformation network's infrastructure is spread across Europe, report says, https://therecord.media/doppelganger-disinformation-infrastructure-european-companies

9. Disinformation, Malware and Drugs: Aeza's cyber crime portfolio \- Qurium Media Foundation, https://www.qurium.org/press-releases/aeza-disinformation-and-drugs/

10. Threat Intelligence Report: The SDA / Structura / Doppelgänger, Influence Operations, Infrastructure, Reach, and Potential \- DomainTools Investigations, https://dti.domaintools.com/research/sda-structura-doppelganger-influence-ops

11. State-Sponsored Russian Media Leverages Meliorator Software for Foreign Malign Influence Activity, https://www.aha.org/system/files/media/file/2024/07/joint-cybersecurity-advisory-tlp-clear-state-sponsored-russian-media-leverages-meliorator-software-for-foreign-malign-influence-activity-7-9-2024.pdf

12. Report: West manipulates media to cast its influence \- Chinadaily.com.cn, https://global.chinadaily.com.cn/a/202209/13/WS631f67fda310fd2b29e772b9.html

13. Russian Interference in Israel's Internal Affairs \- JISS, https://jiss.org.il/en/rakov-russian-interference-in-israels-internal-affairs/

14. (U) The Psychology of (Dis)information: Case Studies and Implications \- CNA.org., https://www.cna.org/reports/2021/10/The-Psychology-of-%28Dis%29information-Case-Studies-and-Implications.pdf

15. Fewer Bots, More Ads: The Pentagon's Evolving Online Influence Campaigns | Lawfare, https://www.lawfaremedia.org/article/fewer-bots--more-ads--the-pentagon-s-evolving-online-influence-campaigns

16. THE SPIRIT OF LIBERTY: AT HOME, IN THE WORLD \- Imgix, https://gwbcenter.imgix.net/Publications/Resources/gwbi-the-spirit-of-liberty-full-report.pdf

17. policy blueprint for countering authoritarian interference in democracies \- German Marshall Fund, https://www.gmfus.org/sites/default/files/Policy-Blueprint.pdf

18. Russian state media uses covert software for disinformation bots, US warns, https://therecord.media/russia-disinformation-bots-social-media-us-canada-netherlands-alert

19. 1 National Security Archive Cyber Vault A CHRONOLOGY OF THE CYBER ASPECTS OF THE WAR IN UKRAINE1 (As of March 27, 2023\) PRE-202, https://nsarchive.gwu.edu/sites/default/files/2023-04/Ukraine%20cyber%20chronology%20-%20as%20of%20March%2027%2C%202023.pdf

20. Burned After Reading: Endless Mayfly's Ephemeral Disinformation Campaign, https://citizenlab.ca/research/burned-after-reading-endless-mayflys-ephemeral-disinformation-campaign/

21. A Russian Bot Farm Used AI to Lie to Americans. What Now? \- CSIS, https://www.csis.org/analysis/russian-bot-farm-used-ai-lie-americans-what-now

22. Unheard Voice: Evaluating five years of pro-Western covert influence operations \- OCR of the Document | National Security Archive, https://nsarchive.gwu.edu/media/29432/ocr

23. open hearing on foreign influence operations' use of social media platforms (third party expert, https://www.intelligence.senate.gov/wp-content/uploads/2024/08/sites-default-files-hearings-chrg-115shrg30959.pdf

24. Defending American Democracy in the Post-Truth Age: A Roadmap to a Whole-of-Society Approach \- DTIC, https://apps.dtic.mil/sti/trecms/pdf/AD1126516.pdf

25. Censorship Industrial Complex \- Document Repository, https://docs.house.gov/meetings/IF/IF16/20230328/115561/HHRG-118-IF16-20230328-SD012.pdf

26. MC Weekly Update 1/30: No One Expects the Copyright Order \- Stanford Law School, https://law.stanford.edu/podcast/mc-weekly-update-1-30-no-one-expects-the-copyright-order/

27. US, international authorities seize Russian AI bot farm | CyberScoop, https://cyberscoop.com/us-international-authorities-seize-russian-ai-bot-farm/

28. C2PA and SynthID in OpenAI-generated images, https://help.openai.com/en/articles/8912793-c2pa-and-synthid-in-openai-generated-images

29. C2PA Security Considerations, https://spec.c2pa.org/specifications/specifications/2.4/security/Security\_Considerations.html

30. Why Instagram Changed the 'Made with AI' Label to 'AI Info' | Release \#253 | Magazine, https://www.cizucu.com/en/magazines/2024-07-ai-info

31. Meta changes name of 'Made with AI' label for AI-generated content to 'AI info' after photographer complaints \- GIGAZINE, https://gigazine.net/gsc\_news/en/20240702-meta-ai-info/

32. Our Approach to Labeling AI-Generated Content and Manipulated Media \- About Meta, https://about.fb.com/news/2024/04/metas-approach-to-labeling-ai-generated-content-and-manipulated-media/

33. Meta Backtracks On Made With AI Labels For Real Photos Following Angry Backlash, https://hothardware.com/news/meta-backtracks-made-with-ai

34. Meta Is Changing Artificial Intelligence Labels After Real Photos Were Marked As AI \- CNET, https://www.cnet.com/tech/services-and-software/meta-is-changing-artificial-intelligence-labels-after-real-photos-were-marked-as-ai/

35. Meta set to alter labels for AI-generated images | The Jerusalem Post, https://www.jpost.com/business-and-innovation/article-809048

36. Meta changes Made with AI label to AI Info on photographers' images after backlash, https://www.indiatoday.in/technology/news/story/meta-changes-made-with-ai-label-to-ai-info-on-photographers-images-after-backlash-2561087-2024-07-02

37. AI Content Moderation Trends for 2026 | Blog \- Conectys, https://www.conectys.com/blog/posts/ai-content-moderation-trends-for-2026/

38. Content Moderation | AI4media, https://www.ai4media.eu/wp-content/uploads/2023/12/Factsheet\_AI-and-Content-Moderation.pdf

39. Lawfare Daily: Making Sense of the Doppelganger Disinformation Operation, with Thomas Rid, https://www.lawfaremedia.org/article/lawfare-daily--making-sense-of-the-doppelganger-disinformation-operation--with-thomas-rid

40. Inside Doppelganger – How Russia uses EU companies for its propaganda \- CORRECTIV, https://correctiv.org/en/fact-checking-en/2024/07/22/inside-doppelganger-how-russia-uses-eu-companies-for-its-propaganda/

41. Spamouflage \- Wikipedia, https://en.wikipedia.org/wiki/Spamouflage

42. Just So You Know: Foreign Threat Actors Likely to Use a Variety of Tactics to Develop and Spread Disinformation During 2024 U.S. General Election Cycle \- IC3, https://www.ic3.gov/PSA/2024/PSA241018

43. ChinaAngVirus disinformation campaign \- Wikipedia, https://en.wikipedia.org/wiki/ChinaAngVirus\_disinformation\_campaign

44. Pentagon not denying report on discrediting of Chinese vaccines \- Philippine News Agency, https://www.pna.gov.ph/articles/1227042

45. International relations in public health: the Pentagon's anti-vax campaign during COVID-19 pandemic \- Oxford Academic, https://academic.oup.com/jpubhealth/article/46/4/e685/7718885

46. U.S. Disinformation Plan Put Geopolitics Above Global Health, https://www.thinkglobalhealth.org/article/us-disinformation-plan-put-geopolitics-above-global-health

47. Massive secret network was pushing Western narrative \- China Daily, https://www.chinadailyhk.com/hk/article/288869

48. 1 National Security Archive Cyber Vault A CHRONOLOGY OF THE CYBER ASPECTS OF THE WAR IN UKRAINE1 (Through March 12, 2024\) PRE-2, https://nsarchive.gwu.edu/sites/default/files/2024-03/Ukraine%20cyber%20chronology%20-%20as%20of%20March%2012,%202024.pdf

49. How Generative AI Is Transforming Journalism: Development, Application and Ethics \- MDPI, https://www.mdpi.com/2673-5172/5/2/39

50. Trolls in Slovakian election tap AI deepfakes to spread disinformation | The Straits Times, https://www.straitstimes.com/world/europe/trolls-in-slovakian-election-tap-ai-deepfakes-to-spread-disinformation

51. Nunes memo \- Wikipedia, https://en.wikipedia.org/wiki/Nunes\_memo

52. Justice Dept. Disrupts Russian Influence Campaign, Indicts Russian Nationals | Lawfare, https://www.lawfaremedia.org/article/justice-dept.-disrupts-russian-influence-campaign--indicts-russian-nationals

53. Pentagon Stands by Secret Anti-Vaccination Disinformation Campaign in Philippines After Reuters Report \- Military.com, https://www.military.com/daily-news/2024/06/14/pentagon-stands-secret-anti-vaccination-disinformation-campaign-philippines-after-reuters-report.html

54. Artificial Intelligence and the Transformation of the Media System \- MDPI, https://www.mdpi.com/2673-8392/6/2/45

55. Media Literacy, https://www.isbe.net/Documents/Media-Literacy-Fact-Sheet.pdf

56. Illinois' Pioneering Media Literacy Program Lacks Resources, Oversight \- Pulitzer Center, https://pulitzercenter.org/stories/illinois-pioneering-media-literacy-program-lacks-resources-oversight

57. English Language Arts and Media Services \- J Sterling Morton High School District 201, https://www.morton201.org/morton201academics/general-academics/english-department-overview

58. J. Sterling Morton High School District 201, https://resources.finalsite.net/images/v1723488798/morton201org/oschx98zasdvksvc4gbd/20242025StudentHandbook.pdf

59. District Facts \- J Sterling Morton High School District 201, https://www.morton201.org/district-info/district-facts

60. Career and Technical Education \- J Sterling Morton High School District 201, https://www.morton201.org/morton201academics/general-academics/career-and-technical-education

61. How Can Solutions-Focused Researchers Scale Their Work Beyond the University?, https://impact.stanford.edu/article/how-can-solutions-focused-researchers-scale-their-work-beyond-university