[Paper Review] Trust, Accountability, and Autonomy in Knowledge Graph-Based AI for Self-Determination

Luis Ibáñez, John Domingue+4 more|arXiv (Cornell University)|Jun 30, 2023

Legal and Policy Issues2 citations

TL;DR

This paper proposes a research agenda for Knowledge Graph (KG)-based AI that supports citizen self-determination through trust, accountability, and autonomy. It integrates neuro-symbolic AI with decentralized infrastructure, explainable AI, and machine-readable policies to enable transparent, auditable, and user-controlled AI systems, positioning KGs as foundational for ethical, human-centered AI in society.

ABSTRACT

Knowledge Graphs (KGs) have emerged as fundamental platforms for powering intelligent decision-making and a wide range of Artificial Intelligence (AI) services across major corporations such as Google, Walmart, and AirBnb. KGs complement Machine Learning (ML) algorithms by providing data context and semantics, thereby enabling further inference and question-answering capabilities. The integration of KGs with neuronal learning (e.g., Large Language Models (LLMs)) is currently a topic of active research, commonly named neuro-symbolic AI. Despite the numerous benefits that can be accomplished with KG-based AI, its growing ubiquity within online services may result in the loss of self-determination for citizens as a fundamental societal issue. The more we rely on these technologies, which are often centralised, the less citizens will be able to determine their own destinies. To counter this threat, AI regulation, such as the European Union (EU) AI Act, is being proposed in certain regions. The regulation sets what technologists need to do, leading to questions concerning How the output of AI systems can be trusted? What is needed to ensure that the data fuelling and the inner workings of these artefacts are transparent? How can AI be made accountable for its decision-making? This paper conceptualises the foundational topics and research pillars to support KG-based AI for self-determination. Drawing upon this conceptual framework, challenges and opportunities for citizen self-determination are illustrated and analysed in a real-world scenario. As a result, we propose a research agenda aimed at accomplishing the recommended objectives.

Motivation & Objective

• To address the threat of centralized AI systems undermining citizen self-determination due to opaque, unaccountable data processing.
• To identify core research pillars—trust, accountability, and autonomy—that can guide the ethical development of KG-based AI.
• To propose a conceptual framework integrating machine-readable norms, decentralized infrastructure, and explainable neuro-symbolic AI for ethical AI design.
• To analyze real-world challenges in AI governance and data control through a self-determination lens.
• To establish a research agenda that aligns technological development with fundamental human rights and EU AI Act principles.

Proposed method

• Proposes a conceptual framework where trust, accountability, and autonomy are the three pillars for KG-based AI systems.
• Integrates machine-readable norms and policies as foundational components for specifying and enforcing ethical AI behavior.
• Advocates for decentralized infrastructure and decentralized KG management to reduce reliance on centralized platforms.
• Promotes explainable neuro-symbolic AI by combining symbolic knowledge graphs with neural models (e.g., LLMs) for interpretable reasoning.
• Leverages standards from the Semantic Web (RDF, OWL) and decentralized identity systems (DIDs, Verifiable Credentials) for data and identity integrity.
• Uses real-world scenarios to illustrate challenges in data control, transparency, and compliance, informing the research agenda.

Experimental results

Research questions

• RQ1 How can KG-based AI systems be designed to support individual data self-determination in the face of centralized data control?
• RQ2 What technical and architectural components are required to ensure accountability and transparency in KG-driven AI decision-making?
• RQ3 How can explainable neuro-symbolic AI be implemented to provide justifications for AI outputs while preserving semantic integrity?
• RQ4 What role do decentralized infrastructures and machine-readable policies play in enabling user autonomy over personal data?
• RQ5 How can existing regulatory frameworks like the EU AI Act be technically operationalized in KG-based AI systems?

Key findings

• The integration of KGs with large language models (LLMs) enables more interpretable and context-aware AI, supporting explainable decision-making.
• Decentralized infrastructure and decentralized KG management are essential to prevent monopolistic control and enhance user autonomy.
• Machine-readable policies and norms can be used to encode legal and ethical requirements, enabling automated compliance checking in AI systems.
• Current AI systems often lack transparency and accountability, especially when data and models are hidden behind corporate firewalls.
• The proposed research agenda identifies actionable technical pathways to align KG-based AI with human rights, particularly self-determination.
• The vision of an 'Oh yeah?' button for trust justification remains unrealized in mainstream browsers, highlighting the need for technical and policy innovation in AI transparency.