Back to search

NAERINGSPH-Nærings-phd

On Bots, Brokers, and the Affordances of AI

Alternative title: Maskiner, Meglere og Muligheter

Awarded: NOK 1.6 mill.

The R&D project examines the organizational uptake of data-driven and algorithmic decision-making (ADM), with a focus on datafication in the setting of global maritime trade. ADM here refers to data captured through digital devices and processed by algorithms where the goal is to predict an object's class or (future) behavior based on its current or past behavior. Despite increased adoption of AI/ML-based technologies, research suggest that increased data visibility and use of automated decision-making can have unintended consequences including distortion of meaning and biased inference. While research suggest that the inclusion of human oversight and control might provide a solution to these challenges, however, there exists little knowledge of how organizations can balance human-interpretable systems and automated performance in this context. The purpose of the project, thus, is to explore and document new, dynamic forms of human-algorithm collaboration, how data can be leveraged as a strategic asset, and knowledge about their consequences for organizations and society.

Modellene som er utviklet i prosjektet kan brukes metodisk av organisasjoner for å strukturere oppgavefordeling mellom mennesker og algoritmer med tanke på å optimalisere for kvalitet, nøyaktighet og effektivitet i det de innfører og utvikler intelligente algoritmer og data-drevne prosesser i virksomheten. Resultatene kan bidra til å utvikle et mer nyansert syn på anskaffelse av data og intelligente algoritmer, og med dette bringe med seg økt oppmerksomhet om viktigheten av å håndtere usikkerheten knyttet til data og slik unngå unødvendige investeringer. Resultatene har ledet til økt samarbeid mellom bedriften, en av bedriftens kunder og UoH-institusjonen, deriblant en nyoppstartet prosess for utvikling av et nytt forskningsprosjekt, økt tverrfaglig samarbeid på tvers av kundevirksomheten, forbedret innsikt i og organisering av arbeidsprosesser og roller i møte med datafisering og opptak av intelligente algoritmer.

Advances in digital technologies and ever growing amounts of data outpace organizations' ability to attend to and make sense of such vast data in knowledge work. Consequently, many organizations face uncertainty in decision-making and problem-solving, which places special demands on ICT design. To cope with such knowledge problems, more and more organizations turn to intelligent technologies, which promise a major breakthrough in how organizations collect, analyze, and act on data and information. Reportedly, 78 percent of managers will trust algorithmic advice in making decisions. Yet old problems loom large. Evidence of fatal failures and repeat problems of control in adaptation of artificial intelligence (AI)-based technologies abound in press and research alike. Since no set of permanent solutions exist, such phenomena call for more research into the sociotechnical arrangements of algorithms, humans, and organizational practices. This project takes an insider view from behind the scenes at a multinational firm in international maritime trade and its dealings with the design and adaptation of data-driven intelligent technologies. We take an interpretive, pragmatist perspective and draw on practice-based approaches to uncover the sociotechnical practices and processes which AI spring from. In particular, we are interested to develop a detailed understanding of emergent configurations of human-machine collaboration and feedback mechanisms as a source for data-driven strategic advantage. Despite much scholarly interest in both AI and the role of uncertainty, surprisingly little attention has been paid to the role of emerging data-driven and algorithmic technologies in shaping firms' knowledge practices and strategies under conditions of uncertainty.

Publications from Cristin

No publications found

No publications found

No publications found

No publications found

Funding scheme:

NAERINGSPH-Nærings-phd