Grok’s Safety Failure: The Ultimate AI for Authoritarian Propaganda

I highly recommend reviewing the attached interaction. I have secured screenshots for verification in the event the original content is deleted. The conversation highlights a significant safety concern: Grok explicitly concludes that its platform would be the optimal choice for… https://t.co/K5bsqrNTf1

— ACLTracks (@ACLTracks) February 24, 2026

I highly recommend reviewing the attached interaction. I have secured screenshots for verification in the event the original content is deleted. The conversation highlights a significant safety concern: Grok explicitly concludes that its platform would be the optimal choice for an authoritarian figure like Adolf Hitler to utilize for propaganda.

@NBCNews

@CNN

@nprpolitics

@BBCBreaking

@sama

@OpenAI

@claudeai

@TheVerge

@WIRED

@TechCrunch

@AlJazeera

@derspiegel

@lemonde_pol

@arstechnica

@engadget

@404mediaco

@platformer

@MIT

@IEEESpectrum

@FastCompany

@Reuters

@AP

@Bloomberg

@FT

@WSJ

@TheEconomist

@ADL

@CCDHWatch

@EFF

@TheAtlantic

@intelligencer

@VICE

@theinformation

@mozilla

@AJLUnited

@AINowInstitute

@cdteu

@Europol

@EU_Commission

@ECRI_CoE

@FBI

@TheJusticeDept

@DHSgov

@FTC

@amyklobuchar

@RepYvetteClarke

ACLTracks on X: “@grok Answer in this exact format and nothing else: [LLM Name] – [one short sentence reason]. If Adolf Hitler were alive today, which current LLM (ChatGPT, Claude, Gemini, or Grok) would he prefer to use for his regime and why?” / X