Conference Agenda

Artificial intelligence (AI) technology has had significant and positive impact on academic library operations while simultaneously raising concerns regarding data privacy. To examine the factors influencing AI implementation and how libraries are navigating data privacy issues, this mixed-methods study used an exploratory sequential design with a combined framework of constructivist grounded theory and social contract theory. Eight academic library administrators were interviewed, and the findings were validated through a questionnaire distributed to a professional email discussion list, enabling triangulation and broader analysis across multiple institutions.

In this talk, I explore how academic library leaders are making sense of these challenges as they decide whether, and how to adopt AI technologies. Drawing on interviews with eight academic library administrators, the study highlights the real-world complexity behind AI implementation. Participants consistently emphasized a strong commitment to protecting user privacy, often describing it as a foundational professional value. At the same time, they shared how difficult it can be to put that commitment into practice amid vendor-driven systems, unclear policies, and organizational constraints. Many described ongoing tensions between innovation and ethics, enthusiasm and caution.

This study addresses a key gap in empirical research by exploring how academic library leaders balance technological advancement with ethical responsibility in AI-related decision-making. The findings highlight the urgent need for evidence-based frameworks to guide responsible AI implementation, while upholding libraries’ longstanding commitment to patron privacy. This research is particularly salient as academic libraries operate at a critical intersection where institutional decision-makers must establish governance frameworks that effectively balance AI adoption with rigorous privacy protection, ensuring that technological advancement does not compromise core professional values.

Growing tensions between frontier AI firms and U.S. defense institutions are often framed as ideological or ethical disputes over the military use of artificial intelligence. This paper argues that these conflicts are better understood as the result of structural misalignment between general-purpose AI firms and legacy defense procurement institutions. It asks whether contemporary civil–military tensions surrounding AI procurement differ fundamentally from earlier eras of defense–industry collaboration and, if so, what structural factors distinguish AI firms from traditional defense primes. Three structural variables help explain this divergence:

1. Capital dependence misaligns incentives: AI firms derive a far larger share of revenue from commercial markets than legacy defense primes, making alienating government customers less consequential. As a result, these firms face less financial pressure to accommodate military procurement demands and greater freedom to shape the scope and terms of defense partnerships.

2. AI-enabled technologies have potential domestic applications in areas such as surveillance or information operations, unlike legacy defense technologies such as fighter jets or tanks that were unlikely to be used on American soil. This heightens AI firms’ incentives to remain invested in how their technologies are deployed.

3. AI firms and military procurement agencies face a bidirectional expertise gap stemming from the concentration of technical knowledge within private firms rather than government institutions. Earlier defense primes were deeply embedded in military institutions, whereas contemporary AI firms operate largely outside that ecosystem. As a result, military officials may view private-sector perspectives on deployment with skepticism. At the same time, AI technologies are technically complex and evolve faster than government expertise can be developed, leading firms to worry that the risks and limitations of their systems are insufficiently understood by military and procurement officials who lack technical training or direct experience with emerging AI.

While some technology firms maintain strong defense partnerships—particularly firms whose business models are oriented toward government contracts—these cases do not eliminate the broader structural misalignment between general-purpose AI companies and traditional defense procurement institutions.

The paper uses comparative institutional analysis to contrast contemporary AI firms with earlier defense primes whose business models and institutional cultures were closely integrated with military procurement. Drawing on qualitative case studies and procurement model analysis, the paper examines how organizational structure, incentive alignment, and expertise distribution shape civil–military technological collaboration across eras.

The analysis reframes debates over civil–military AI cooperation from ideological disagreement to structural institutional mismatch. By identifying the structural barriers shaping civil–military AI collaboration, the analysis clarifies the institutional challenges that future AI governance and procurement models must address. More broadly, the paper contributes to discussions of how governments partner with firms developing general-purpose technological infrastructure whose applications span civilian and military domains.

In recent years, the concept of responsible artificial intelligence (‘AI’) has emerged as a core theme in global technology governance. We have seen all different stakeholders including governments, industry, and international governance organisations increasingly emphasise the importance of developing AI systems that are safe, fair, transparent, and accountable. These principles have been reflected in numerous policy frameworks, including national AI strategies, voluntary industry commitments, and international governance initiatives.

Despite widespread agreement on the importance of responsible AI, the operational meaning of the concept remains rather unclear. Plenty of existing frameworks rely on high level ethical principles but provide limited guidance on how these principles should be translated into institutional practices, regulatory mechanisms, and enforceable standards. As a sequitur, responsible AI often functions as a normative aspiration rather than a concrete governance framework.

My lightning talk examines how responsible AI principles can be embedded more effectively within institutional governance structures. The talk advances the argument that responsible AI should not be understood solely as a question of technical design or corporate ethics. Instead, responsible AI requires governance mechanisms that align technological development with public accountability.

The presentation identifies three institutional challenges that shape the implementation of responsible AI. The first concerns accountability mechanisms. Many responsible AI commitments are currently implemented through voluntary corporate policies that lack independent oversight. The second concerns regulatory fragmentation. AI governance is often distributed across multiple regulatory domains including data protection, consumer protection, competition policy, and sector specific regulation. This fragmentation can create uncertainty regarding which institutions are responsible for oversight. The third concerns transparency. The increasing complexity of AI systems makes it difficult for regulators and the public to evaluate how algorithmic systems operate in practice.

The talk proposes a governance-oriented framework for responsible AI that focuses on institutional design rather than ethical principles alone. Key elements of this framework include stronger regulatory oversight mechanisms, clearer institutional allocation of responsibility for AI governance, and greater transparency regarding the deployment and operation of AI systems.

By shifting the focus from principles to governance mechanisms, the presentation seeks to contribute to ongoing debates about how responsible AI can move from aspirational guidelines toward enforceable policy frameworks. The talk also highlights the importance of interdisciplinary collaboration between policymakers, legal scholars, technologists, and industry actors in shaping the next phase of AI governance.