/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

OpenAI says it disrupted 10 malicious operations in the past three months, four of which likely originated in China and targeted multiple countries and topics

Chinese propagandists are using ChatGPT to write posts and comments on social media sites — and also to create performance reviews detailing …

NPR Shannon Bond

Context & Ripple Effects

This sits in a growing record of AI providers identifying influence-related misuse of generative tools: OpenAI had previously reported disrupting five covert influence operations seeking to shape public opinion or political outcomes.

The cases broaden the pattern beyond a single country or electoral target. Subsequent coverage of Iranian accounts removed over a US-election influence effort and China-linked activity points to repeated attempts to use model access for scalable persuasion work.

First-order effects

  • The identified operations lose access to ChatGPT for drafting social posts, comments, and related campaign materials, interrupting the workflows OpenAI detected.
  • OpenAI must treat influence operations as an active abuse category across multiple countries and topics, rather than a narrowly election-specific risk.

Second-order effects

  • Other model providers and social platforms face greater pressure to detect coordinated misuse that begins with AI-generated copy but is distributed through external accounts.
  • The repeated disclosures make account enforcement and abuse reporting more consequential parts of a model provider's trust and safety operation, alongside model-level refusals.

Third-order effects

  • If these cases continue, frontier-model access is likely to become a more explicit point of control in cross-border information operations, linking AI safety enforcement to geopolitical risk management.
  • The durable challenge will be operational: providers can disrupt known accounts, but the corpus shows influence actors can adapt targets, narratives, and platforms, making prevention an ongoing monitoring problem.

The trend: Generative-AI providers are becoming active gatekeepers against the use of model access to scale covert influence campaigns.

Discussion

  • @kjephd Kevin Elliott on bluesky
    I'm sorry, but this is hilarious: Chinese spies are using ChatGPT to do their disinfo homework for them, and to write reports to their bosses  —  www.npr.org/2025/06/05/n...  [image]
  • @justinhendrix Justin Hendrix on bluesky
    Interesting to see OpenAI adopt the reporting style of major platforms for state-sponsored campaigns.  [embedded post]