AiOfAi: 4th Workshop on Adverse Impacts and Collateral Effects of Artificial Intelligence Technologies 

We invite AI researchers and practitioners across different disciplines and knowledge backgrounds to submit contributions dealing with the following (or related) topics:

• Adverse Applications of Generative AI and Machine Learning (ML):

  • • AI-driven online deception, misinformation/disinformation, and social manipulation

    • Jailbreaks, prompt/data injection, and scalable content automation using LLMs

    • Fraud, impersonation, and social engineering attacks powered by generative models 

• Mitigation Strategies and Safety Mechanisms for AI/ML:

  • • Threat modeling, risk assessment, and adversarial testing of ML systems

    • Explainable AI (XAI), human-in-the-loop systems, and oversight frameworks

    • Safety guardrails for LLMs and multimodal AI systems

    • Retrieval-Augmented Generation (RAG) reliability and hallucination mitigation

    • Behavioral modeling and intrusion detection

    • Privileged Access Management (PAM) and privilege creep detection 

• Evaluation and Oversight of Generative and Agentic AI Systems:

  • • Containment, alignment, and interpretability of generative models

    • Benchmarking, scenario libraries, and simulation testbeds

    • Post-deployment monitoring, incident reporting, and documentation

    • Secure agent architectures and compliance in high-risk AI domains 

• Technical and Societal Impacts of AI and ML Systems: 

  • • Privacy breaches, data leakage, and model inversion attacks

    • Bias, discrimination, and representational harms in ML

    • Psychological and societal effects of Human-AI interaction

    • Trust erosion and polarization in digital ecosystems

• Legal, Ethical, and Governance Considerations in AI:

  • •  Intellectual property and content ownership

    • Regulatory compliance for AI systems (e.g., GDPR, AI Act)

    • AI in surveillance and law enforcement

    • Responsible AI, transparency, and auditing mechanisms 

Special topics of interest:

1. Misuse & Societal Harms of Generative AI: Deepfakes and synthetic voice; Misinformation/disinformation; Scam automation and social engineering; Privacy leakage; Bias and representational harms; Accessibility and language equity; Harms in high-risk domains 

2. Safety, Evaluation & Governance of Generative Systems: Jailbreaks, prompt/data injection, containment; hallucination mitigation; RAG reliability; post-deployment monitoring; incident reporting; human oversight; alignment; interpretability; safety cases; auditing; compliance processes for high-risk GenAI. 

We welcome submissions spanning the full range of theoretical and applied work, including user research, methods, datasets, tools, simulations, demos, and practical evaluations. 

Submission Details