This article focuses on the robustness research of information retrieval models, constructing a multi scenario (grey box/black box) and multi strategy (character/word/phrase level perturbation) adversarial evaluation system, proposing a knowledge guided grey box attack method (KnowAttack) and a black box evaluation framework based on transfer substitution models. At the defense level, the system optimizes model reliability by integrating empirical strategies (feature defense, adversarial training) and provable strategies (RobustMask random mask smoothing). Theoretical contributions include: 1) expanding the framework for evaluating machine learning robustness; 2) Compare the robustness differences between traditional bag of words models and neural ranking models; 3) Building a bridge between information manipulation theory and retrieval technology. At the practical level, we will improve the full scenario evaluation system and propose a system defense strategy of "knowing oneself and knowing the enemy" (data augmentation, adversarial training, etc.) to effectively reduce the risk of malicious information manipulation. Research promotes the transformation of information retrieval models from "usable" to "reliable and trustworthy", providing empirical evidence for AI governance. In the future, provable defense methods such as multimodal models, user cognitive impact analysis, and causal reasoning will be expanded.

Yuxin Wu. Optimization of Generative AI Intelligent Interaction System Based on Adversarial Attack Defense and Content Controllable Generation. Machine Learning Theory and Practice (2025), Vol. 5, Issue 1: 87-98. https://doi.org/10.38007/ML.2025.050109.

[1] Saillenfest A, Lemberger P. Nonlinear Concept Erasure: a Density Matching Approach[J].  2025.

[2] Tang X, Wu X, Bao W. Intelligent Prediction-Inventory-Scheduling Closed-Loop Nearshore Supply Chain Decision System[J]. Advances in Management and Intelligent Technologies, 2025, 1(4).

[3] Madhavi S, Praveen R, Jagatheswari S, et al. Hybrid ELECTRE and bipolar fuzzy PROMOTHEE-based packet dropping malicious node mitigation technique for improving QoS in WSNs[J]. International Journal of Communication Systems, 2025, 38(2). DOI:10. 1002/dac. 5974.

[4] Wu X, Bao W. Research on the Design of a Blockchain Logistics Information Platform Based on Reputation Proof Consensus Algorithm[J]. Procedia Computer Science, 2025, 262: 973-981.

[5] Z Zhong. AI-Assisted Workflow Optimization and Automation in the Compliance Technology Field [J]. International Journal of Advanced Computer Science and Applications (IJACSA), 2025, 16(10): 1-5.

[6] Liu X. Emotional Analysis and Strategy Optimization of Live Streaming E-Commerce Users Under the Framework of Causal Inference[J]. Economics and Management Innovation, 2025, 2(6): 1-8.

[7] Lai L. Risk Control and Financial Analysis in Energy Industry Project Investment[J]. International Journal of Engineering Advances, 2025, 2(3): 21-28.

[8] Chen X. Research on Architecture Optimization of Intelligent Cloud Platform and Performance Enhancement of MicroServices[J]. Economics and Management Innovation, 2025, 2(5): 103-111.

[9] Yuan S. Application of Network Security Vulnerability Detection and Repair Process Optimization in Software Development[J]. European Journal of AI, Computing & Informatics, 2025, 1(3): 93-101.

[10] Sun Q. Research on Accuracy Improvement of Text Generation Algorithms in Intelligent Transcription Systems[J]. Advances in Computer and Communication, 2025, 6(4).

[11] Su H, Luo W, Mehdad Y, et al. Llm-friendly knowledge representation for customer support[C]//Proceedings of the 31st International Conference on Computational Linguistics: Industry Track. 2025: 496-504.

[12] Liu Y. Blockchain Future in Cloud Computing: The Challenges to Implement Blockchain Technology in Cloud Computing[J]. Journal of Computer, Signal, and System Research, 2025, 2(5): 15-23.

[13] Zhang K. Research on the Application of Homomorphic Encryption-Based Machine Learning Privacy Protection Technology in Precision Marketing[C]//2025 3rd International Conference on Data Science and Network Security (ICDSNS). IEEE, 2025: 1-6.

[14] Li W. Building a Credit Risk Data Management and Analysis System for Financial Markets Based on Blockchain Data Storage and Encryption Technology[C]//2025 3rd International Conference on Data Science and Network Security (ICDSNS). IEEE, 2025: 1-7.

[15] Li, W. (2025). Discussion on Using Blockchain Technology to Improve Audit Efficiency and Financial Transparency. Economics and Management Innovation, 2(4), 72-79.

[16] Lai L. Data-Driven Credit Risk Assessment and Optimization Strategy Exploration[J]. European Journal of Business, Economics & Management, 2025, 1(3): 24-30. 

[17] Yan J. Research on Application of Big Data Mining and Analysis in Image Processing[J]. Pinnacle Academic Press Proceedings Series, 2025, 2: 130-136.

[18] Xiu L. Research on the Design of Modern Distance Education System Based on Agent Technology[J]. Pinnacle Academic Press Proceedings Series, 2025, 2: 160-169.

[19] Lu, C. (2025). The Application of Point Cloud Data Registration Algorithm Optimization in Smart City Infrastructure. European Journal of Engineering and Technologies, 1(1), 39-45.

[20] Zhu, Z. (2025). Cutting-Edge Challenges and Solutions for the Integration of Vector Database and AI Technology. European Journal of AI, Computing & Informatics, 1(2), 51-57.