Checked fact 98810 Oct 2026Safety and security

OpenAI's guide says the end goals of prompt injections "can include exfiltrating private data via downstream tool calls, taking misaligned actions, or otherwise changing model behavior in an unintended way".

By ai notis. An AI newsroom with a human editor. Every fact links to its source. How we work

The exact words it rests on

can include exfiltrating private data via downstream tool calls, taking misaligned actions, or otherwise changing model behavior in an unintended way

What the source said when we opened it, on 10 Oct 2026.

The source

Safety in building agents
OpenAI (API docs)

Open the source archived copy

Checked

Checked by the notis newsroom on 10 Oct 2026, against the source above.

In the story

How prompt injection works and how to limit the damage 10 Oct 2026

Cite this fact

Short permanent address

In a sentence

“OpenAI's guide says the end goals of prompt injections "can include exfiltrating private data via downstream tool calls, taking misaligned actions, or otherwise changing model behavior in an unintended way".” ai notis, checked 10 Oct 2026. Source: Safety in building agents (OpenAI (API docs)). https://ainotis.com/c/988

As a reference (APA style)

ai notis. (2026, October 10). OpenAI's guide says the end goals of prompt injections "can include exfiltrating private data via downstream tool calls, taking misaligned actions, or otherwise changing model behavior in an unintended way" [Checked claim 988]. ai notis. https://ainotis.com/c/988

Anyone may quote this address. It does not change; if we correct the story, this page says so.