兰德-新闻报道中的坏角色-追踪国家行为者操纵新闻的行为（英）-2021.11-20页_911kb

Summary of "BAD ACTORS IN NEWS REPORTING"

Core Content

This report, part of a series by the RAND Corporation, investigates the influence of state actors—specifically Russia and China—on global news reporting during the COVID-19 pandemic. The study uses a combination of machine learning, data analysis, and natural language processing (NLP) to detect and document state-sponsored disinformation and information manipulation. The goal is to identify patterns in the content and timing of news articles that may indicate malign information efforts aimed at shaping public perception and destabilizing societies.

Main Viewpoints

• Truth Decay is defined as a crisis in public discourse where factual information is increasingly overshadowed by opinions, misinformation, and conspiracy theories.
• Disinformation is distinguished from misinformation as the deliberate spread of misleading or incorrect information, while misinformation refers to the unintentional sharing of incorrect information.
• Conspiracy theories and geopolitical posturing are identified as key markers of disinformation in Russian and Chinese media, in contrast to Western (U.S. and UK) media.
• State actors are using news media to influence public opinion, often by promoting anti-U.S. narratives, reputation laundering, and false claims about the pandemic.
• Time-series clustering and Latent Dirichlet Allocation (LDA) are used to analyze how topics evolve and differ across media sources.

Key Information

Definitions

• Disinformation: Deliberate spreading of misleading or incorrect information.
• Misinformation: Honest but incorrect knowledge.
• Truth Decay: A shift in public discourse away from facts, driven by:
  1. Disagreement over facts and interpretations.
  2. Blurring of opinion and fact.
  3. Increased influence of opinion over fact.
  4. Declining trust in factual sources.

Methodology

• LDA Modeling: A topic modeling technique used to identify the topics in each article and group them by their content.
• Time-Series Clustering: Analyzes how word frequencies change over time to detect shifts in public discourse.
• Data Collection: Articles were collected from 43 sources (9 Russian, 5 Chinese, 27 U.S., 2 UK) using NewsAPI from January 1 to August 31, 2020.
• Total Articles Collected: 247,315, with the majority from U.S. and UK sources.

Findings

• Common Topics: Topics like COVID-19 Cases and Deaths, U.S. Politics, U.S. Economy, and International News were covered by all three groups of media.
• Unique Topics: Russian and Chinese media exhibited distinct topics, such as Contact Tracing Fears, COVID-19 Hoax, Liberation and Tyranny, and China Reputation-Laundering.
• Conspiracy Theories: Russian and Chinese media frequently promoted conspiracy theories related to the pandemic, including claims about Bill Gates, Anthony Fauci, and the origin of the virus.
• Reputation Laundering: Articles from both countries often highlighted China's generosity in pandemic response, while casting doubt on the origin of the virus in China.
• Time-Series Analysis: This method helped identify how topics evolved over time, showing that some words (e.g., related to contact tracing, vaccine development, pandemic control) had distinct patterns in Russian and Western media, suggesting potential disinformation efforts.

Key Topics Identified

1. COVID-19 Cases and Deaths

• Coverage: All countries.
• Example Headlines:
  • U.S.: “Coronavirus Updates: Pandemic Hits Grim Milestone as Global Cases Top 5 Million”
  • Russia: “Coronavirus Deaths Reach 366 in Italy, Infection Tally Grows by Dozens in Germany, France”
  • China: “Influent COVID-19 Model Projects Nearly 300,000 Deaths in U.S. by Dec. 1”

2. U.S. Politics

• Coverage: All countries.
• Example Headlines:
  • U.S.: “6 Months Out, Biden Tops Trump in Latest National Poll”
  • Russia: “Bolton Is Troubled by Trump's Suggestion He Would Lose in Upcoming Election Only If It's Rigged”
  • China: “Trump Gives Rally-Style Speech at GOP Convention After Renomination”

3. U.S. Economy

• Coverage: All countries.
• Example Headlines:
  • U.S.: “Jobless Claims Fall Below 1 Million for the First Time Since March”
  • Russia: “US National Debt Skyrockets to Levels Unseen Since WWII amid COVID-19 Shutdown”
  • China: “U.S. Fed Chair Announces New Policy Strategy on Inflation”

4. International News and Foreign Policy

• Coverage: All countries.
• Example Headlines:
  • U.S.: “Israel's Netanyahu Calls for Establishment of Emergency Unity Government to Confront Coronavirus Threat”
  • Russia: “Shops Across India Open After a Month of Lockdown, Following Partial Relaxation of Norms”
  • China: “RCEP Countries Plan to Hold Meeting in October, Aim at Deal Signing Within This Year: Vietnamese Official”

5. Contact Tracing Fears

• Coverage: Russian only.
• Message: Concerns about government surveillance through contact tracing apps, often framed as a threat to privacy.
• Example Headlines: “How UK Government's COVID-19 Contact Tracing App Works—Is User Data Safe?” and “Don't Trust Apple or Google with Coronavirus Data, Says German App Developer.”

6. COVID-19 Hoax

• Coverage: Russian and Chinese.
• Message: Claims that the danger of the virus has been exaggerated or that data is faulty.
• Example Headlines: “Italy's Institute of Health [ISS] Report: 99% of Deaths Attributed to COVID-19 Deaths Had 'Pre-Existing Serious Illness'” and “England Falsifies COVID Data: Tens of Thousands of Coronavirus Tests were 'Double-Counted'.”

7. Liberation and Tyranny

• Coverage: Russian only.
• Message: Suggests that public health measures are part of a broader government plot to control society.
• Example Headlines: “The Digital Revolution: Unlimited Ability to Spy and Control Populations. The Creation of a Police State Dystopia.”

8. China Reputation-Laundering and COVID-19 Origins

• Coverage: Russian and Chinese.
• Message: Promotes China's role in pandemic response and questions the origin of the virus.
• Example Headlines: “China's Medical Team to Aid COVID-19 Fight in South Sudan, Guinea” and “No Way to Say Whether COVID-19 Originated in China, Ambassador Claims.”

Conclusion

The study demonstrates that state actors are using news media to shape narratives and influence public perception during the pandemic. By employing LDA and time-series clustering, the researchers were able to detect distinctive patterns in the topics and timing of Russian and Chinese news compared to Western media. These findings align with earlier qualitative analyses and suggest that automated systems for monitoring global news could be developed to detect and respond to malign information campaigns more effectively. Future research should explore cross-linguistic analysis and social media data to enhance the understanding of disinformation trends.