The Balsillie School of International Affairs (BSIA) is a centre for advanced research and teaching on global governance and international public policy, located in Waterloo, Ontario. As one of the largest social sciences initiatives in Canada, the school is a collaborative partnership between the University of Waterloo, Wilfrid Laurier University, and the Centre for International Governance Innovation. The BSIA is an affiliate member of the Association of Professional Schools of International Affairs, a group of schools that educate leaders in international affairs. The BSIA is housed in the north and west wings of the CIGI Campus. Admission to BSIA is highly selective.
Industry 5.0, deploying artificial intelligence (AI) at its core, reframes industrial evolution from a predominantly technology- and efficiency-driven innovation model toward a virtuously human-centric, sustainable, and resilient model of value creation by organizations. This review paper, based on an interdisciplinary literature review, explores how AI, within the Industry 5.0 paradigm, reshapes economic logics, the understanding of information asymmetry, and sustainability trajectories, and the implications for entrepreneurial strategy and business model innovation, which demand the development of a new form of organizational intelligence. While the literature suggests that AI, when deployed within a mature Industry 5.0 framework, could generate synergistic economic and sustainability values through circular, human-centered, and digitally augmented systems, human–AI co-intelligence gains are contingent on insights that address systems quality, reskilling, ethics, and reorienting resources from overly short-term profit maximization toward wisdom for long-term socio-ecological, climate resilience, and ESG performance. This study introduces a framework for tackling organizational sustainability dynamics, anticipating the emergence of new industries and the retransformation of enduring ones amid creative destruction in the AI era. Future studies to fill knowledge gaps and implications for human competencies that will enhance organizational intelligence are articulated.
Abstract Sensitivity is invoked by diverse actors in contexts of ecological struggle, and it plays important roles in how these conflicts are framed and waged. However, there has not yet been a systematic analysis of its complex meanings and uses. This article analyzes how “sensitivity” is mobilized in four key eco-political discourses. Western scientists and policymakers use the term “sensitivity” to describe certain ecosystems as particularly responsive to disruptions, or unable to regulate themselves in the face of disruptions that compromise their productivity, functionality, or value. Western medical discourses use norms of sensitivity to environmental stimuli to designate and discipline “normal” versus “abnormal” bodies and to enforce racial and gendered hierarchies. Meanwhile, states (along with police, military, and corporate actors) inflame and engender sensitivities in targeted bodies and ecosystems to wield and expand power and territorial domination. They do so by creating unequal structural distributions of stimuli and through the weaponization of particular human and nonhuman sensitivities. However, diverse communities affected by eco-political violence also cultivate and mobilize sensitivities in complex ways—for instance, by drawing on overlapping sensitivities to foster solidarities with harmed ecosystems, and by embracing bodily sensitivities as sources of critical knowledge and collective protection. Drawing on environmental justice, disability justice, and liberatory disability discourses, and the broader knowledge traditions in which they are grounded (including Black and Indigenous feminisms and queer theory), this article plumbs the multivalent uses of “sensitivity” at sites of environmental violence and its potential to contribute to struggles for eco-political justice.
Federated unlearning (FU) aims to remove adversarial contributions from distributed learning systems without retraining, a capability relevant to AI-driven energy systems under data poisoning. Existing FU methods are evaluated almost entirely at the model level, implicitly assuming that clearing adversarial influence from parameters restores downstream system behavior. We test that assumption in agentic energy systems, where decisions are sequential, memory-dependent, and coupled to physical state. This study does not implement or benchmark a specific client-level federated-unlearning algorithm; retraining, clean fine-tuning, and checkpoint rollback are evaluated as controlled FU-inspired model-recovery proxies. We propose a system-level, agent-based evaluation framework that captures the interaction among predictive models, memory-aware decision agents, and dynamic system states, and use it to evaluate these proxies under cyber and physical perturbations. Across ten seeds, Proxy_Retrain and Proxy_Finetune produced mean paired-ASR values numerically close to the clean mean, while the full-horizon aggregate positive-imbalance (API) proxy stayed above the clean baseline for every arm and no proxy improved it relative to the poisoned run. Restricted to the post-recovery window, mean absolute imbalance did not differ detectably from clean, so these data show a model-level/system-level divergence over the full horizon but do not establish persistent post-recovery trajectory deviation. All findings are limited to this stylized simulation, ten seeds, the evaluated parameters, and the three recovery proxies; no real grid, utility deployment, or production FU implementation was tested.
