A5067816317- Misinformation and Its Impacts
- Vaccine Coverage and Hesitancy
- Social Media and Politics
- Ethics and Social Impacts of AI
- Terrorism, Counterterrorism, and Political Violence
- Ethics and Legal Issues in Pediatric Healthcare
- SARS-CoV-2 and COVID-19 Research
- Topic Modeling
- Media Studies and Communication
- Hate Speech and Cyberbullying Detection
- Psychology of Moral and Emotional Judgment
- Child and Adolescent Health
- Middle East and Rwanda Conflicts
- Cybersecurity and Cyber Warfare Studies
- COVID-19 epidemiological studies
- Global Peace and Security Dynamics
- International Relations and Foreign Policy
- Global Energy Security and Policy
- Economic Sanctions and International Relations
- Global Security and Public Health
- Data-Driven Disease Surveillance
- Focus Groups and Qualitative Methods
- Chinese history and philosophy
- Sexuality, Behavior, and Technology
- Socioeconomic Development in MENA
Stanford University
2020-2024
Generative language models have improved drastically, and can now produce realistic text outputs that are difficult to distinguish from human-written content. For malicious actors, these bring the promise of automating creation convincing misleading for use in influence operations. This report assesses how might change operations future, what steps be taken mitigate this threat. We lay out possible changes behaviors, content online operations, provide a framework stages model-to-influence...
The 2020 US election was accompanied by an effort to spread a false meta-narrative of widespread voter fraud. This took hold among substantial portion the population, undermining trust in procedures and results, eventually motivating events 6 January 2021. We examine this as domestic participatory disinformation campaign which variety influencers—including hyperpartisan media political operatives—worked alongside ordinary people produce amplify misleading claims, often unwittingly. To better...
The emergence of tools based on Large Language Models (LLMs), such as OpenAI's ChatGPT, Microsoft's Bing Chat, and Google's Bard, has garnered immense public attention. These incredibly useful, natural-sounding mark significant advances in natural language generation, yet they exhibit a propensity to generate false, erroneous, or misleading content -- commonly referred "hallucinations." Moreover, LLMs can be exploited for malicious applications, generating false but credible-sounding...
Russian influence operations on social media have received significant attention following the 2016 US presidential elections. Here, scholarship has largely focused covert strategies of Russia-based Internet Research Agency and overt Russia's largest international broadcaster RT (Russia Today). But since 2017, a number new news providers linked to state emerged, less research these channels how they may support contemporary operations. We conduct qualitative content analysis 2,014 Facebook...
To gain insight into how Chinese state media is communicating about the coronavirus pandemic to outside world, we analyzed a collection of posts from their English-language presence on Facebook. We observed three recurring behaviors: sharing positive stories and promoting Communist Party's (CCP) response, rewriting recent history in manner favorable
Much of the research and discourse on risks from artificial intelligence (AI) image generators, such as DALL-E Midjourney, has centered around whether they could be used to inject false information into political discourse. We show that spammers scammers—seemingly motivated by profit or clout, not ideology—are already using AI-generated images gain significant traction Facebook. At times, Facebook Feed is recommending unlabeled users who neither follow Pages posting nor realize are...
Anonymity is an important principle online. However, malicious actors have long used misleading identities to conduct fraud, spread disinformation, and carry out other deceptive schemes. With the advent of increasingly capable AI, bad can amplify potential scale effectiveness their operations, intensifying challenge balancing anonymity trustworthiness In this paper, we analyze value a new tool address challenge: "personhood credentials" (PHCs), digital credentials that empower users...
Most content consumed online is curated by proprietary algorithms deployed social media platforms and search engines. In this article, we explore the interplay between these human agency. Specifically, consider extent of entanglement or coupling humans along a continuum from implicit to explicit demand. We emphasize that interactions people have with not only shape users’ experiences in moment but because mutually shaping nature such systems can also longer-term effects through modifications...
When governments run influence operations they may leverage in-house capabilities, outsource to digital mercenaries, or use a combination of these strategies. We theorize that because it provides plausible deniability if the operation is uncovered, and offers access cutting-edge tactics beyond those common established government institutions. Using data from Facebook, Twitter, Instagram, we test implications this theory via two covert online campaign case studies, each focused on Syria,...
Researchers of foreign and domestic influence operations document tactics that frequently recur in covert propaganda campaigns on social media, including backstopping fake personas with plausible biographies or histories, using GAN-generated images as profile photos, outsourcing account management to paid organizations. These tactics, however, can be applied outside the political realm. In this paper, we describe how these three are leveraged serve a commercial purpose: lead generation for...
This study introduces methods for identifying search directives — content that could prompt an online and explores their presence on social media. Search can be effective tool indirect influence, because instead of guiding people directly to (e.g., a news article), they indirectly guide it through independent intermediary Google Search). By directing viewers "do own research" known, trusted, or seemingly objective intermediary, have advantages in terms both persuasion, by making feel as if...