Influence activity
Coordinated efforts to shape public opinion or political outcomes, often using deceptive or manipulative means.
Influence activity is when groups try to change what people think or how they act, especially on a large scale. They often use fake information or misleading tactics to achieve their goals. This can be done to affect elections, public opinion on issues, or to spread propaganda.
For example, a group might use AI to create many fake social media accounts that all post similar messages supporting a particular political candidate. These messages might seem like they come from real people, but they are part of a coordinated effort to sway voters.
Influence activity refers to coordinated campaigns designed to manipulate public opinion, political discourse, or societal attitudes. These operations often leverage deceptive tactics, including the creation of inauthentic personas and the dissemination of tailored disinformation, to achieve specific strategic objectives. The scale and sophistication of these activities have increased with the advent of advanced AI tools that can automate content generation and amplify messaging across platforms.
Frontier AI labs are increasingly identifying and disrupting influence activities powered by their own models. Recent reports highlight operations using AI for foreign propaganda, election interference, and social engineering. This shows a growing concern about the misuse of AI for malicious influence campaigns.
- Operation “Fish Food”: Russia-origin content farm activityOpenAI · OpenAI News ·
- Operation “Stop News”: Recidivist influence activityOpenAI · OpenAI News ·
- Operation “VAGue Focus”: Social engineering and influence activityOpenAI · OpenAI News ·
- Operation “Sponsored Discontent”: Influence activityOpenAI · OpenAI News ·
- Operation “STORM-2035”: Iran-origin influence activityOpenAI · OpenAI News ·
- Operation "Spamouflage": China-linked influence activityOpenAI · OpenAI News ·
- OpenAI ChatGPT
- OpenAI actively monitors and bans accounts engaged in influence activity, particularly when AI is used to generate content for these operations.
- Disinformation
- Disinformation is false information spread deliberately to decei
The term 'influence activity' is widely used in cybersecurity and political science to describe efforts to shape perceptions and behaviors. Its application in the context of AI misuse has become more prominent as AI tools are increasingly employed in these campaigns.
You can observe how AI can be used to generate persuasive text by asking a model to write a short social media post arguing for or against a fictional policy. Then, try asking it to write similar posts from different viewpoints to see how easily content can be generated for varied messaging.
The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.
- PromptRole-play a student struggling with math for AI tutor simulationSimulate a student's math struggles and uncertainty for AI tutor evaluation. · 5 min
- TemplateReusable workflow templates for healthcare teamsTeams can create and share templates for common workflows like documentation or prior authorizations. · 15 min
- TemplateTemplate for building an AI tool tailored to your workA structured approach to defining and building a custom AI tool for specific business needs. · 15 min
- TemplateEvals CSV prompt list for agent skillsA CSV file template for listing prompts used in evaluating agent skills. · 20 min
- HeuristicDouble coding agent task completion with docs linksProviding coding agents with a documentation link significantly increases their task completion rate. · 2 min
Questions people ask
- How is influence activity different from just spreading fake news?
- Influence activity is a coordinated campaign to shape opinions, which may include spreading fake news. It's about the organized effort and intent, not just the content itself.
- Can AI make influence activity more dangerous?
- Yes, AI can create content at scale, personalize messages, and automate the spread of information, making influence campaigns more effective and harder to detect.
First seen on Subvolts May 1, 2024. Updated Sep 9, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.