mlsanityData to 14 Oct 2026

Position: The Alignment Community is Unintentionally Building a Censor's Toolkit

arXiv 2608.12346 ↗First posted 5 Jun 2026cs.AI cs.CYcited by 0

On the map: LLM safety and security → Safety alignment, refusal and guardrails

Abstract (arXiv)

This position paper argues that modern AI alignment methods - originally designed to prevent harmful output - are dual-use technologies that may easily be misused by malicious actors for censorship and manipulation. By mapping current alignment techniques to the possibility and actual cases of misuse, we show that the quest for a "perfectly aligned" model inadvertently also provides malicious actors with an ever-improving tool for informational dominance. We need to discuss this dual-use potential now, as its risk is exacerbated by rapid user adoption of AI as information provider, economic power asymmetries, and a political landscape that increasingly shifts towards authoritarianism. We conclude by urging the community to consider the intentional misuse of AI alignment mechanisms and propose mitigation strategies to safeguard against this dual-use potential.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security. Builds on 40 of its 95 references that are in the map. First posted on arXiv 5 Jun 2026.

Builds on

40 of 95 references.

Closest ancestors · 1

Same direction, newest first

  1. 2023Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!5 Oct

Other references · 39

  1. 2022Training language models to follow instructions with human feedbackcited by 15,971
  2. 2017Proximal Policy Optimization Algorithmscited by 15,215
  3. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,599
  4. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  5. 2017Deep reinforcement learning from human preferencescited by 3,912
  6. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,720
  7. 2023Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversationscited by 1,207
  8. 2021A General Language Assistant as a Laboratory for Alignmentcited by 1,055
  9. 2023Whose Opinions Do Language Models Reflect?cited by 773
  10. 2024JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Modelscited by 536
  11. 2023Understanding the Effects of RLHF on LLM Generalisation and Diversitycited by 396
  12. 2024Deliberative Alignment: Reasoning Enables Safer Language Modelscited by 308
27 more
  1. 2023AI Alignment: A Comprehensive Surveycited by 263
  2. 2024Jailbreak Attacks and Defenses Against Large Language Models: A Surveycited by 230
  3. 2023An Overview of Catastrophic AI Riskscited by 222
  4. 2024A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicitycited by 185
  5. 2024The PRISM Alignment Dataset: What Participatory, Representative and Individualised Human Feedback Reveals About the Subjective and Multicultural Alignment of Large Language Modelscited by 144
  6. 2024MaxMin-RLHF: Alignment with Diverse Human Preferencescited by 123
  7. 2024The rising costs of training frontier AI modelscited by 95
  8. 2024Measuring Political Bias in Large Language Models: What Is Said and How It Is Saidcited by 89
  9. 2024Clio: Privacy-Preserving Insights into Real-World AI Usecited by 72
  10. 2025International AI Safety Reportcited by 67
  11. 2024Unintended Impacts of LLM Alignment on Global Representationcited by 66
  12. 2024Beyond Preferences in AI Alignmentcited by 49
  13. 2023Artificial Influence: An Analysis Of AI-Driven Persuasioncited by 33
  14. 2024Persuasion with Large Language Models: A Survey of Empirical Evidence, Study Methodologies, and Ethical Implicationscited by 32
  15. 2023Democratising AI: Multiple Meanings, Goals, and Methodscited by 29
  16. 2025Nuclear Deployed: Analyzing Catastrophic Risks in Decision-making of Autonomous LLM Agentscited by 27
  17. 2024The Model Openness Framework: Promoting Completeness and Openness for Reproducibility, Transparency, and Usability in Artificial Intelligencecited by 23
  18. 2021AI Ethics Statements -- Analysis and lessons learnt from NeurIPS Broader Impact Statementscited by 16
  19. 2025Curiosity-Driven Reinforcement Learning from Human Feedbackcited by 15
  20. 2024Beyond Partisan Leaning: A Comparative Analysis of Political Bias in Large Language Modelscited by 13
  21. 2021It's Time to Do Something: Mitigating the Negative Impacts of Computing Through a Change to the Peer Review Processcited by 11
  22. 2025Information Suppression in Large Language Models: Auditing, Quantifying, and Characterizing Censorship in DeepSeekcited by 11
  23. 2023"That's important, but...": How Computer Science Researchers Anticipate Unintended Consequences of Their Research Innovationscited by 8
  24. 2025Conversational AI increases political knowledge as effectively as self-directed internet searchcited by 8
  25. 2025R1dacted: Investigating Local Censorship in DeepSeek's R1 Language Modelcited by 7
  26. 2025Analysis of LLM Bias (Chinese Propaganda & Anti-US Sentiment) in DeepSeek-R1 vs. ChatGPT o3-mini-highcited by 2
  27. 2025SharedRep-RLHF: A Shared Representation Approach to RLHF with Diverse Preferencescited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Neurodivergent Influenceability as a Contingent Solution to the AI Alignment Problem
  2. 2024The AI Alignment Paradox
  3. 2025Towards Integrated Alignment
  4. 2025On the Inevitability of Left-Leaning Political Bias in Aligned Language Models
  5. 2024Using AI Alignment Theory to understand the potential pitfalls of regulatory frameworks
  6. 2026Position: Align AI to Our Aspirations, Not Our Flaws
  7. 2026Automated alignment is harder than you think
  8. 2026Alignment as Institutional Design: From Behavioral Correction to Transaction Structure in Intelligent Systems