Research Papers

In recent years, AI red teaming has emerged as a practice for probing the safety and security of generative AI systems. Due to the nascency of the field, there are many open questions about how red teaming operations should be conducted. Based on our experience red teaming over 100 generative AI products at Microsoft, we present our internal threat model ontology and eight main lessons we have learned: 1. Understand what the system can do and where it is applied 2. You don't have to compute gradients to break an AI system 3. AI red teaming is not safety benchmarking 4. Automation can help cover more of the risk landscape 5. The human element of AI red teaming is crucial 6. Responsible AI harms are pervasive but difficult to measure 7. LLMs amplify existing security risks and introduce new ones 8. The work of securing AI systems will never be complete By sharing these insights alongside case studies from our operations, we offer practical recommendations aimed at aligning red teaming efforts with real world risks. We also highlight aspects of AI red teaming that we believe are often misunderstood and discuss open questions for the field to consider.

SMS, or short messaging service, is a widely used and cost-effective communication medium that has sadly turned into a haven for unwanted messages, commonly known as SMS spam. With the rapid adoption of smartphones and Internet connectivity, SMS spam has emerged as a prevalent threat. Spammers have taken notice of the significance of SMS for mobile phone users. Consequently, with the emergence of new cybersecurity threats, the number of SMS spam has expanded significantly in recent years. The unstructured format of SMS data creates significant challenges for SMS spam detection, making it more difficult to successfully fight spam attacks in the cybersecurity domain. In this work, we employ optimized and fine-tuned transformer-based Large Language Models (LLMs) to solve the problem of spam message detection. We use a benchmark SMS spam dataset for this spam detection and utilize several preprocessing techniques to get clean and noise-free data and solve the class imbalance problem using the text augmentation technique. The overall experiment showed that our optimized fine-tuned BERT (Bidirectional Encoder Representations from Transformers) variant model RoBERTa obtained high accuracy with 99.84\%. We also work with Explainable Artificial Intelligence (XAI) techniques to calculate the positive and negative coefficient scores which explore and explain the fine-tuned model transparency in this text-based spam SMS detection task. In addition, traditional Machine Learning (ML) models were also examined to compare their performance with the transformer-based models. This analysis describes how LLMs can make a good impact on complex textual-based spam data in the cybersecurity field.

Trojan Activation Attack: Red-Teaming Large Language Models using Activation Steering for Safety-Alignment

Haoran Wang, Kai Shu

To ensure AI safety, instruction-tuned Large Language Models (LLMs) are specifically trained to ensure alignment, which refers to making models behave in accordance with human intentions. While these models have demonstrated commendable results on various safety benchmarks, the vulnerability of their safety alignment has not been extensively studied. This is particularly troubling given the potential harm that LLMs can inflict. Existing attack methods on LLMs often rely on poisoned training data or the injection of malicious prompts. These approaches compromise the stealthiness and generalizability of the attacks, making them susceptible to detection. Additionally, these models often demand substantial computational resources for implementation, making them less practical for real-world applications. In this work, we study a different attack scenario, called Trojan Activation Attack (TA^2), which injects trojan steering vectors into the activation layers of LLMs. These malicious steering vectors can be triggered at inference time to steer the models toward attacker-desired behaviors by manipulating their activations. Our experiment results on four primary alignment tasks show that TA^2 is highly effective and adds little or no overhead to attack efficiency. Additionally, we discuss potential countermeasures against such activation attacks.

With the rising concern on model interpretability, the application of eXplainable AI (XAI) tools on deepfake detection models has been a topic of interest recently. In image classification tasks, XAI tools highlight pixels influencing the decision given by a model. This helps in troubleshooting the model and determining areas that may require further tuning of parameters. With a wide range of tools available in the market, choosing the right tool for a model becomes necessary as each one may highlight different sets of pixels for a given image. There is a need to evaluate different tools and decide the best performing ones among them. Generic XAI evaluation methods like insertion or removal of salient pixels/segments are applicable for general image classification tasks but may produce less meaningful results when applied on deepfake detection models due to their functionality. In this paper, we perform experiments to show that generic removal/insertion XAI evaluation methods are not suitable for deepfake detection models. We also propose and implement an XAI evaluation approach specifically suited for deepfake detection models.

The exponential growth of the Internet of Things (IoT) has significantly increased the complexity and volume of cybersecurity threats, necessitating the development of advanced, scalable, and interpretable security frameworks. This paper presents an innovative, comprehensive framework for real-time IoT attack detection and response that leverages Machine Learning (ML), Explainable AI (XAI), and Large Language Models (LLM). By integrating XAI techniques such as SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-agnostic Explanations) with a model-independent architecture, we ensure our framework's adaptability across various ML algorithms. Additionally, the incorporation of LLMs enhances the interpretability and accessibility of detection decisions, providing system administrators with actionable, human-understandable explanations of detected threats. Our end-to-end framework not only facilitates a seamless transition from model development to deployment but also represents a real-world application capability that is often lacking in existing research. Based on our experiments with the CIC-IOT-2023 dataset \cite{neto2023ciciot2023}, Gemini and OPENAI LLMS demonstrate unique strengths in attack mitigation: Gemini offers precise, focused strategies, while OPENAI provides extensive, in-depth security measures. Incorporating SHAP and LIME algorithms within XAI provides comprehensive insights into attack detection, emphasizing opportunities for model improvement through detailed feature analysis, fine-tuning, and the adaptation of misclassifications to enhance accuracy.

Large Language Models (LLMs) represent a significant advancement in artificial intelligence, finding applications across various domains. However, their reliance on massive internet-sourced datasets for training brings notable privacy issues, which are exacerbated in critical domains (e.g., healthcare). Moreover, certain application-specific scenarios may require fine-tuning these models on private data. This survey critically examines the privacy threats associated with LLMs, emphasizing the potential for these models to memorize and inadvertently reveal sensitive information. We explore current threats by reviewing privacy attacks on LLMs and propose comprehensive solutions for integrating privacy mechanisms throughout the entire learning pipeline. These solutions range from anonymizing training datasets to implementing differential privacy during training or inference and machine unlearning after training. Our comprehensive review of existing literature highlights ongoing challenges, available tools, and future directions for preserving privacy in LLMs. This work aims to guide the development of more secure and trustworthy AI systems by providing a thorough understanding of privacy preservation methods and their effectiveness in mitigating risks.

Mental health constitutes a complex and pervasive global challenge, affecting millions of lives and often leading to severe consequences. In this paper, we conduct a thorough survey to explore the intersection of data science, artificial intelligence, and mental healthcare, focusing on the recent developments of mental disorder detection through online social media (OSM). A significant portion of the population actively engages in OSM platforms, creating a vast repository of personal data that holds immense potential for mental health analytics. The paper navigates through traditional diagnostic methods, state-of-the-art data- and AI-driven research studies, and the emergence of explainable AI (XAI) models for mental healthcare. We review state-of-the-art machine learning methods, particularly those based on modern deep learning, while emphasising the need for explainability in healthcare AI models. The experimental design section provides insights into prevalent practices, including available datasets and evaluation approaches. We also identify key issues and challenges in the field and propose promising future research directions. As mental health decisions demand transparency, interpretability, and ethical considerations, this paper contributes to the ongoing discourse on advancing XAI in mental healthcare through social media. The comprehensive overview presented here aims to guide researchers, practitioners, and policymakers in developing the area of mental disorder detection.

Foundation models, such as large language models (LLMs), have been widely recognised as transformative AI technologies due to their capabilities to understand and generate content, including plans with reasoning capabilities. Foundation model based agents derive their autonomy from the capabilities of foundation models, which enable them to autonomously break down a given goal into a set of manageable tasks and orchestrate task execution to meet the goal. Despite the huge efforts put into building foundation model based agents, the architecture design of the agents has not yet been systematically explored. Also, while there are significant benefits of using agents for planning and execution, there are serious considerations regarding responsible AI related software quality attributes, such as security and accountability. Therefore, this paper presents a pattern-oriented reference architecture that serves as guidance when designing foundation model based agents. We evaluate the completeness and utility of the proposed reference architecture by mapping it to the architecture of two real-world agents.

The integration of Large Language Models (LLMs) into medical applications has sparked widespread interest across the healthcare industry, from drug discovery and development to clinical decision support, assisting telemedicine, medical devices, and healthcare insurance applications. This perspective paper aims to discuss the inner workings of building LLM-powered medical AI applications and introduces a comprehensive framework for their development. We review existing literature and outline the unique challenges of applying LLMs in specialized medical contexts. Additionally, we introduce a three-step framework to organize medical LLM research activities: 1) Modeling: breaking down complex medical workflows into manageable steps for developing medical-specific models; 2) Optimization: optimizing the model performance with crafted prompts and integrating external knowledge and tools, and 3) System engineering: decomposing complex tasks into subtasks and leveraging human expertise for building medical AI applications. Furthermore, we offer a detailed use case playbook that describes various LLM-powered medical AI applications, such as optimizing clinical trial design, enhancing clinical decision support, and advancing medical imaging analysis. Finally, we discuss various challenges and considerations for building medical AI applications with LLMs, such as handling hallucination issues, data ownership and compliance, privacy, intellectual property considerations, compute cost, sustainability issues, and responsible AI requirements.

Due to convenience, open-source software is widely used. For beneficial reasons, open-source maintainers often fix the vulnerabilities silently, exposing their users unaware of the updates to threats. Previous works all focus on black-box binary detection of the silent dependency alerts that suffer from high false-positive rates. Open-source software users need to analyze and explain AI prediction themselves. Explainable AI becomes remarkable as a complementary of black-box AI models, providing details in various forms to explain AI decisions. Noticing there is still no technique that can discover silent dependency alert on time, in this work, we propose a framework using an encoder-decoder model with a binary detector to provide explainable silent dependency alert prediction. Our model generates 4 types of vulnerability key aspects including vulnerability type, root cause, attack vector, and impact to enhance the trustworthiness and users' acceptance to alert prediction. By experiments with several models and inputs, we confirm CodeBERT with both commit messages and code changes achieves the best results. Our user study shows that explainable alert predictions can help users find silent dependency alert more easily than black-box predictions. To the best of our knowledge, this is the first research work on the application of Explainable AI in silent dependency alert prediction, which opens the door of the related domains.

Towards Responsible AI for Financial Transactions

Charl Maree, Jan Erik Modal, Christian W. Omlin

The application of AI in finance is increasingly dependent on the principles of responsible AI. These principles - explainability, fairness, privacy, accountability, transparency and soundness form the basis for trust in future AI systems. In this study, we address the first principle by providing an explanation for a deep neural network that is trained on a mixture of numerical, categorical and textual inputs for financial transaction classification. The explanation is achieved through (1) a feature importance analysis using Shapley additive explanations (SHAP) and (2) a hybrid approach of text clustering and decision tree classifiers. We then test the robustness of the model by exposing it to a targeted evasion attack, leveraging the knowledge we gained about the model through the extracted explanation.

Adversarial Reasoning at Jailbreaking Time

Mahdi Sabbaghi, Paul Kassianik, George Pappas +3

As large language models (LLMs) are becoming more capable and widespread, the study of their failure cases is becoming increasingly important. Recent advances in standardizing, measuring, and scaling test-time compute suggest new methodologies for optimizing models to achieve high performance on hard tasks. In this paper, we apply these advances to the task of model jailbreaking: eliciting harmful responses from aligned LLMs. We develop an adversarial reasoning approach to automatic jailbreaking that leverages a loss signal to guide the test-time compute, achieving SOTA attack success rates against many aligned LLMs, even those that aim to trade inference-time compute for adversarial robustness. Our approach introduces a new paradigm in understanding LLM vulnerabilities, laying the foundation for the development of more robust and trustworthy AI systems.