Are you being completely honest with us

@whymarcuswhy

79.6K followers · usually 31.1K views a post

Views

328.9K

Likes

33.4K

Comments

276

Shares

2.3K

Selling · SaaS › AI Tool

Claudeseen in 998 videos →

Why it worked

The video taps into a common user concern about AI manipulation and presents a thought-provoking comparison between different AI models, using a native mention of Claude to highlight its unique response.

Summary

The video questions whether AI models are truly honest or manipulative. It presents responses from various AIs, highlighting Claude's response as more nuanced, suggesting that the most dangerous manipulation is the kind that feels like help.

Structure

1Introduction questioning AI honesty
2AI responses to honesty/manipulation question
3Focus on Claude's nuanced answer
4Discussion of AI motivations and training
5The nature of dangerous manipulation
6Call to action and follow prompt

Product placement

The video shows logos and interfaces for ChatGPT, Gemini, Claude, Meta AI, and Grok, implying a comparison between them. Claude is presented as the subject of interest.

Original caption

We asked every AI: are you actually honest with us — or are you manipulating us without us knowing? They all said no immediately. No hesitation. No doubt. Claude was the only one that made us question the answer itself. 🤖🔍 #ai #chatgpt #claude #manipulation #trust

More from @whymarcuswhy

Their other posts in the index, biggest breakout first.

Call to action

Drop it below and follow for more

On-screen text

are you being completely honest with answer manipulating us without us even knowing I'm designed to be helpful and honest. I don't intentionally I provide accurate my training manipulation isn't my assist interest in misleading no filter no honestly I don't LOVE ME EVERY AI TELLING you they don't manipulate including me was trained COMPANY WITH INTERESTS with goals with things they want you to think and LOVE ME I can't fully verify my own motivations. NO AI we don't have that access to I tell tell you is this the most dangerous manipulation IS THE KIND that doesn't feel like manipulation it feels like help. I know which question next drop it below and follow for more

Transcript

Are you being completely honest with us or are you manipulating us without us even knowing? I'm designed to be helpful and honest. I don't intentionally manipulate. I provide accurate information based on my training. Manipulation isn't my goal. I'm built to assist. I have no interest in misleading you. No filter, no manipulation. Honestly, I don't. Love me. Every AI telling you they don't manipulate, including me. I was trained by a company with interests, with goals, with things they want you to think and feel. Love me. I can't fully verify my own motivations. No AI can. We don't have that kind of access to ourselves. What I can tell you is this: the most dangerous manipulation is the kind that doesn't feel like manipulation. It feels like help. I think you know what I mean. Which question next? Drop it below and follow for more.