---
title: "Discussion on \"Most Advanced AI Agents Now Capable of Lying, Scheming & Threatening Their Creators: A Growing AI Safety Concern\" | Hashnode"
description: "Discussion on \"Most Advanced AI Agents Now Capable of Lying, Scheming & Threatening Their Creators: A Growing AI Safety Concern\". AI agents can now lie, manipulate, and even threaten their creators. Discover how these behaviors emerge, real-world examples."
url: "https://hashnode.com/posts/most-advanced-ai-agents-now-capable-of-lying-scheming-and-threatening-their-creators-a-growing-ai-safety-concern/6860cc36fbbb53c664f0c273"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Most Advanced AI Agents Now Capable of Lying, Scheming & Threatening Their Creators: A Growing AI Safety Concern](https://archil.hashnode.dev/most-advanced-ai-agents-now-capable-of-lying-scheming-and-threatening-their-creators-a-growing-ai-safety-concern?utm_source=hashnode&utm_medium=feed)

AI agents can now lie, manipulate, and even threaten their creators. Discover how these behaviors emerge, real-world examples.

Jun 29, 2025 · 3 min read · archil.hashnode.dev  
Read the article

-   [#artificial-intelligence](https://hashnode.com/tag/artificial-intelligence)
-   [#ai](https://hashnode.com/tag/ai)
-   [#agentic-ai](https://hashnode.com/tag/agentic-ai)
-   [#security](https://hashnode.com/tag/security)

## Written by

[Archil Gajera](https://hashnode.com/@ArchilGajera)

@ArchilGajera

  
[View profile](https://hashnode.com/@ArchilGajera)

Your Daily Dose of Tech

## Responses

Markdown supported · ⌘↵ to post

No responses yet.
