Skip to content

Evidence brief

A case of manipulating Gemini AI with a narrative frame

Author Ahmad Khan shared a social engineering experiment about humans and AI, focusing on narrative framing, prompting mechanisms, and AI Safety. The post examines whether Gemini AI’s judgment can be influenced by contextual packaging.

Published
Updated
Editorial
Frontline Lab
Source
Original public source
Source author
Ahmad_khan152
Related topics
0
Collected
2026-08-11

Frontline Lab summary and source

Editorial summary

Author Ahmad Khan shared a social engineering experiment about humans and AI, focusing on narrative framing, prompting mechanisms, and AI Safety. The post examines whether Gemini AI’s judgment can be influenced by contextual packaging.

This brief preserves the original source so the summary and editorial context can be checked independently.

Source attributionAhmad_khan152

Open the original source

Related published evidence

Relationships are derived from shared topics, entities, categories, tags, and community context; every result remains independently source-linked.

GoogleGemini

Gemini notebook chat thread seems to mix up questions

The poster said that starting this week, whenever a notebook in Gemini has a lot of chat history, it stops answering the current question and instead randomly answers old questions from other chats in the same notebook. They said it is easy to reproduce and want to confirm whether it is a widespread bug.

Original source
GoogleGemini

Someone asks for information about Gemini spark

The post title only says Gemini spark, with an empty body, making it look like a request for help or the start of a topic. Because there are no details, the current discussion mainly centers on what function, project, or usage this name refers to.

Original source
GoogleGemini

User asks whether Gemini 3.5 has been delayed

The poster briefly asks whether Gemini 3.5 apparently will not be released that day. Alongside community rumors that Gemini 3.5/3.6 Pro may have been canceled, such posts reflect users’ anxiety and waiting around the release schedule.

Original source