Responsible AI Agility for Algorithmic Innovation: The case of Autonomous AI-Enabled Drone Ecosystems

Emerging intelligent technologies such as autonomous drones and vehicles increasingly collect and process information, allocate resources, assign tasks, and make decisions (Faraj et al., 2018; Bailey et al., 2022). Powered by the ubiquity of large databases and artificial intelligence, these technologies learn and act autonomously, enabling new possibilities for digital innovation and collaboration within and across organizations. Although emerging intelligent technologies afford distinctive, novel opportunities, they also raise unique multidimensional challenges. This TREO research talk aims to spark rich, generative and interdisciplinary discussions, providing a much-needed liminal space to reflect on the unintended consequences of the advent of the algorithm era. The talk will use the exemplary empirical case of an AI-enabled swarm of autonomous drones intelligently coordinated and controlled through decentralised organizing practices. The drones operate autonomously, continuously gathering and analyzing vast amounts of real-time weather, traffic and noise data to optimally allocate, assign tasks and coordinate last-mile delivery operations directly to consumers. The research talk will challenge the notion that AI inherently enhances organizational intelligence, highlight the biases and ethical issues embedded in machine learning algorithms, and advocate the integration of responsible AI practices. By examining the emergent and unpredictable nature of the empirical phenomenon, the talk seeks to spark a dialogue on how organizations can balance leveraging AI’s potential with safeguarding ethical considerations. Using provocative initial empirical data, this TREO talk will present an initial theoretical framework of organising emerging intelligent technologies responsibly. The talk will then unpack the black box of AI by deconstructing commonly held ontoepistemological assumptions and their implications. Our collective intelligence has come to rely on a representationalist philosophy that begins with a problem (a request) submitted to generative AI tools, that then provide a relevant narrative likely to answer confidently and brilliantly based on a defined language model parameters. However, the view marginalizes, and ignores co-problematization inquiry practices. To conclude, the talk provides implications for artificial organising processes that require critical examination, providing a much needed platform for a flourishing ecosystem of diverse perspectives.