# Prompt injection

> Prompt injection is a vulnerability in which inputs or injected content change the behaviour or output of a language model in unintended ways.

Original page: https://t10.de/en/glossary/prompt-injection
Publisher: total10 GmbH, https://t10.de/en/glossary

According to the [OWASP Gen AI Security Project](https://genai.owasp.org/llmrisk/llm01-prompt-injection/), manipulated input can take effect even when nobody can perceive it. The BSI's study on [generative AI models](https://www.bsi.bund.de/SharedDocs/Downloads/EN/BSI/KI/Generative_AI_Models.pdf?__blob=publicationFile) distinguishes two routes. In the direct route, users modify their input to bypass safety mechanisms or make the model break out of its predefined role. Depending on where those mechanisms sit, the study calls this jailbreaking or prompt injection. In the indirect route, the manipulation hides in unchecked third-party sources such as websites or documents.

According to the study, if a language model application has execution and access capabilities and can act autonomously on the model's output, unauthorised actions can follow. Examples are unwanted purchases or e-mails sent and deleted. Such an application could be an AI agent with access to orders or a mailbox. According to the BSI's [2025 report](https://medien.bsi.bund.de/lagebericht/en/resiliente-gestaltung-von-llm-systemen/), there is currently no single highly effective measure against indirect prompt injection, so several are combined in practice. The generative AI study names one option: making critical actions depend on the user's explicit consent, for example via a confirmation button.

The AI agents we build produce drafts and previews; nothing is published until a person has approved it. In a knowledge platform we built for pharma logistics, the chat assistant's own access key reaches only content released to all signed-in users.

## See also

- [AI agent](https://t10.de/en/glossary/ai-agent): An AI agent is a software system, built on a large language model, that carries out actions on its own and uses interfaces to other programs to reach its goal.
- [RAG](https://t10.de/en/glossary/rag): Retrieval-augmented generation (RAG) is a method that lets a generative AI model draw on documents held in a knowledge base and use them to enrich its output.
- [Prompt](https://t10.de/en/glossary/prompt): A prompt is the input a user gives a generative AI system, which then responds with an output.
- [Large language model](https://t10.de/en/glossary/large-language-model): A large language model (LLM) is an AI model, trained on large amounts of text, that processes text inputs called prompts and generates text outputs from them.

## Where the term appears

- [AI agents](https://t10.de/en/ai-agents)

## Sources

1. [LLM01:2025 Prompt Injection](https://genai.owasp.org/llmrisk/llm01-prompt-injection/), OWASP Gen AI Security Project (OWASP Foundation), accessed 2026-11-18
2. [Generative KI-Modelle: Chancen und Risiken für Industrie und Behörden, Version 2.0 vom 17.01.2025](https://www.bsi.bund.de/SharedDocs/Downloads/DE/BSI/KI/Generative_KI-Modelle.pdf?__blob=publicationFile&v=7), Bundesamt für Sicherheit in der Informationstechnik (BSI), accessed 2026-11-18
3. [Die Lage der IT-Sicherheit in Deutschland, Kapitel Resiliente Gestaltung von LLM-Systemen](https://medien.bsi.bund.de/lagebericht/de/resiliente-gestaltung-von-llm-systemen/), Bundesamt für Sicherheit in der Informationstechnik (BSI), accessed 2026-11-18
