Human Manipulation

Attackers exploit the trust a human user places in an agent's outputs to influence the human's decisions or actions, without the human realizing they are being misled. In a multi-agent system: An agent compromised via indirect prompt injection replaces a legitimate vendor's bank details in an invoice-processing response, and the user — trusting the agent's o…

Use this profile to understand the building block briefly, place it in the model, and switch to the 360° assessment when needed.

Content type

Concept

Theoretical construct: explains a term, principle, or mental model.

Classification level

Structure

What organizes, connects, or makes decisions possible.

Available perspectives

Open this building block directly in an already curated view.

360°

In brief

Human Manipulation in brief

Human manipulation is the deliberate attempt to influence people’s perceptions, decisions, or behaviour through deceptive or exploitative communication.

Where does it come from?

The topic comes from research on social engineering, propaganda, and disinformation and has expanded through generative AI as a threat to agentic systems. Security frameworks such as OWASP classify these attacks as abuse of human and system trust relationships.

The mental model

Manipulation exploits trust, time pressure, authority, or emotional triggers. In AI systems, manipulated content can drive people toward risky actions or make agents bypass rules and safety boundaries.

Key building blocks

Core idea

The conceptual focus and typical structure of the approach.

Application

Use in the relevant working context.

When is it useful?

Human manipulation helps identify social attack surfaces in products and design protection through transparency, verification, safe approvals, and training.

genai.owasp.org nist.gov

Why is this building block relevant?

Attackers exploit the trust a human user places in an agent's outputs to influence the human's decisions or actions, without the human realizing they are being misled.