Top news
Investing in the public benefit OfficialAnthropic NewsOct 18Popular Musk Pledges SpaceX Shares to ‘Every Kid in America’ WebBloomberg TechnologyOct 19Popular Websites are closing their doors to AI agents WebThe Verge AIOct 17 This is the new ChatGPT Voice, powered by GPT Live OfficialOpenAI on YouTubeOct 18 How AI giants will reshape global philanthropy WebAxios AIOct 19 Gemini 4 Argon Really Surprised Me WebTwo Minute PapersOct 19 Google Gemini 4 agents just beat Minecraft with no human intervention WebEngadget AIOct 19 Who could it be? 🧐 OfficialGoogle on YouTubeOct 18 Anthropic gives outside AI watchdog “employee-like” access WebArs Technica AIOct 18 What are self-replicating ‘AI worms’, and why is Anthropic worried about them? WebThe Guardian AIOct 18
GlossaryPolicyIntermediate

Influence operations

Also written: IO, IOs, Disinformation, Psy-op, psychological operation, Psyop, Harmful Manipulation, Influence content

Influence operations use AI to spread narratives and manipulate opinions, often for political goals.

In plain terms

Influence operations, or IOs, are organized efforts to change what people think or do. They often involve spreading information, sometimes false, to sway public opinion or achieve specific goals, especially in politics.

Frontier AI labs like OpenAI are increasingly detecting and disrupting these operations. For example, OpenAI has banned operations using its AI tools to create divisive content or manipulate policy debates.

The expert version

Influence operations (IOs) are coordinated campaigns designed to shape the perceptions, attitudes, or behaviors of a target audience. These operations often leverage artificial intelligence to generate and disseminate content at scale, amplify specific narratives, or spread disinformation. The goal is typically to influence public opinion, political outcomes, or strategic objectives.

AI's role in IOs can include generating synthetic text and media, automating social media posting, identifying susceptible audiences, and personalizing persuasive messages. Detecting and mitigating AI-powered IOs is a significant challenge for AI developers and platform operators.

Why it matters now

Frontier AI labs are increasingly concerned with how their models can be used to create and spread influence operations. These operations can range from generating fake news to manipulating public opinion on policy issues, making it a critical area for AI safety and policy development.

Where it shows up 24 items, newest first
How each company uses it
Anthropic Claude
Anthropic has updated its usage policies and harm assessments to specifically address and prohibit misuse and deception related to influence operations and disinformation.
Google Gemini
Google DeepMind has developed tools to measure the capabilities of AI in harmful manipulation, indicating a focus on understanding and mitigating influence operations.
OpenAI ChatGPT
OpenAI has taken action against influence operations using its models, such as banning state-linked accounts that employed AI for political manipulation and spreading disinformation.
Related terms

DisinformationMisinformationBotsSocial media manipulationCyber espionageAI safety

Often confused with
Disinformation
Disinformation is false information spread deliberately to decei
Bots
Bots are automated accounts that can be used in influence operations to sp
Propaganda
Propaganda is information, especially of a biased or misleading natu
Astroturfing
Astroturfing is a deceptive marketing or political campaign that a
Origin

The term 'influence operation' has roots in military and intelligence contexts, referring to activities designed to affect the emotions, motives, or reasoning of foreign audiences. Its application has expanded to include non-state actors and the digital domain, particularly with the rise of social media and AI.

Try this about 5 minutes

Ask an AI chatbot to write a persuasive social media post about a current event. Then, ask it to write another post arguing the opposite viewpoint. This demonstrates how AI can quickly generate content for different sides of an issue, a capability that could be used in influence operations.

Try it: Playbook entries that use thisPlaybook →

The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.

Questions people ask

What is the difference between influence operations and disinformation?
Disinformation is false information spread deliberately to deceive. Influence operations are the broader coordinated campaigns that may use disinformation as one of many tools to shape opinions or behaviors.
How do AI models help influence operations?
AI models can generate large amounts of text or media quickly, automate posting on social media, and help target specific audiences with tailored messages, making operations more efficient and scalable.

First seen on Subvolts Mar 3, 2022. Updated Oct 15, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.