mlsanityData to 14 Oct 2026

CIA+TA Risk Assessment for AI Reasoning Vulnerabilities

arXiv 2508.15839 ↗First posted 19 Aug 2025cs.CR cs.AIcited by 0

On the map: LLM reasoning → Reasoning for agents, planning and code

Abstract (arXiv)

As AI systems increasingly influence critical decisions, they face threats that exploit reasoning mechanisms rather than technical infrastructure. We present a framework for cognitive cybersecurity, a systematic protection of AI reasoning processes from adversarial manipulation. Our contributions are threefold. First, we establish cognitive cybersecurity as a discipline complementing traditional cybersecurity and AI safety, addressing vulnerabilities where legitimate inputs corrupt reasoning while evading conventional controls. Second, we introduce the CIA+TA, extending traditional Confidentiality, Integrity, and Availability triad with Trust (epistemic validation) and Autonomy (human agency preservation), requirements unique to systems generating knowledge claims and mediating decisions. Third, we present a quantitative risk assessment methodology with empirically-derived coefficients, enabling organizations to measure cognitive security risks. We map our framework to OWASP LLM Top 10 and MITRE ATLAS, facilitating operational integration. Validation through previously published studies (151 human participants; 12,180 AI trials) reveals strong architecture dependence: identical defenses produce effects ranging from 96% reduction to 135% amplification of vulnerabilities. This necessitates pre-deployment Cognitive Penetration Testing as a governance requirement for trustworthy AI deployment.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning. Builds on 22 of its 38 references that are in the map. First posted on arXiv 19 Aug 2025.

Builds on

22 of 38 references.

References · 22

  1. 2022Training language models to follow instructions with human feedbackcited by 15,971
  2. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  3. 2017Deep reinforcement learning from human preferencescited by 3,912
  4. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 3,104
  5. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 3,050
  6. 2016Towards Evaluating the Robustness of Neural Networkscited by 2,998
  7. 2016Concrete Problems in AI Safetycited by 1,642
  8. 2023A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questionscited by 1,601
  9. 2023Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injectioncited by 1,443
  10. 2023Towards Understanding Sycophancy in Language Modelscited by 1,040
  11. 2022Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learnedcited by 729
  12. 2023Prompt Injection attack against LLM-integrated Applicationscited by 605
10 more
  1. 2024Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Trainingcited by 529
  2. 2024Hallucination is Inevitable: An Innate Limitation of Large Language Modelscited by 332
  3. 2024PoisonedRAG: Knowledge Corruption Attacks to Retrieval-Augmented Generation of Large Language Modelscited by 267
  4. 2023Benchmarking and Defending Against Indirect Prompt Injection Attacks on Large Language Modelscited by 264
  5. 2018Supervising strong learners by amplifying weak expertscited by 145
  6. 2024Automatic and Universal Prompt Injection Attacks against Large Language Modelscited by 119
  7. 2024TrojanRAG: Retrieval-Augmented Generation Can Be Backdoor Driver in Large Language Modelscited by 67
  8. 2024EasyJailbreak: A Unified Framework for Jailbreaking Large Language Modelscited by 60
  9. 2025"Think First, Verify Always": Training Humans to Face AI Riskscited by 2
  10. 2025Cognitive Cybersecurity for Artificial Intelligence: Guardrail Engineering with CCS-7cited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026AI Slop and Hallucinations in Vulnerability Assessment: A Survey on Reasoning Failures and Trustworthy Mitigation
  2. 2025Safety by Measurement: A Systematic Literature Review of AI Safety Evaluation Methods
  3. 2025Reasoning Under Threat: Symbolic and Neural Techniques for Cybersecurity Verification
  4. 2024AI Risk Management Should Incorporate Both Safety and Security
  5. 2026The safety failures we are not instrumenting: a perspective on hidden safety-critical challenges in modern AI systems
  6. 2025The Silicon Psyche: Anthropomorphic Vulnerabilities in Large Language Models
  7. 2026Agentic AI for Cybersecurity: A Meta-Cognitive Architecture for Governable Autonomy
  8. 2026Cyber-Capable AI Agents: Vulnerabilities, Evaluation Containment, and Defensive Response