This paper examines how generative artificial intelligence (GenAI) can be designed and embedded into sustainability reporting workflows to support and transform organizational sensemaking. Small and medium-sized enterprises (SMEs) often struggle with complex reporting requirements and limited sustainability expertise, creating a need for digital support systems. Following an Action design research (ADR) approach, we developed and iteratively refined a GenAI-enabled sustainability reporting assistant and evaluated it across multiple organizational cases. Our findings show that GenAI can support traditional sensemaking activities by structuring information, offering contextual guidance, and lowering cognitive barriers. Beyond this supportive role, GenAI can also actively trigger sensemaking by introducing new perspectives, prompting reflection, and facilitating joint human-AI interpretation. At the same time, GenAI outputs can reduce the depth of users’ engagement and shape which sustainability framings prevail. We contribute to digital sustainability and sensemaking support system literature by deriving design principles for GenAI-enabled sustainability reporting assistants and demonstrating their influence on sustainability reporting practices. The study offers practical implications for integrating GenAI into reporting workflows and highlights avenues for future research.
The spread of fake news poses significant challenges across social, political, and economic domains. Existing countermeasures, like media literacy and fact-checking, have shown limited effectiveness due to user skepticism and cognitive biases. To address this issue, we introduce decision support traceability (DST), which integrates system traceability (explainable AI) and object traceability (chain of sources) to improve users’ ability to assess information credibility. System traceability enhances transparency by explaining AI-generated classifications, while object traceability enables users to track news articles’ origin and evolution. Prior research has examined these mechanisms separately, but their combined effect on disinformation detection remains underexplored. We hypothesize that their integration results in superior performance in identifying fake news. Using design science research (DSR) methodology, we developed and tested a prototype incorporating DST. A controlled online experiment compared three treatments: explainable AI, chain of sources, and their combination. Results show that while both mechanisms independently improve disinformation detection, their combination yields the highest accuracy, demonstrating the complementary nature of DST. This study advances information systems research by empirically validating DST as a unified framework for supporting news credibility assessments. The findings provide actionable design knowledge, emphasizing the need for AI transparency and provenance-tracking features to combat digital disinformation.
As the creator economy rapidly expands into decentralized ecosystems, an increasing number of digital artists are creating their own NFTs on the blockchain. Which characteristics of those creators affect buyers’ valuation of their artworks? The current research identifies one such characteristic: creators’ liking variety, which is exhibited as the diversity in the NFTs they like. Based on 23,579 bidding records collected from Rarible between November 17, 2020, and May 17, 2022, we find that buyers’ willingness to pay is higher for NFTs from creators with higher liking variety, as reflected in the greater diversity of liked NFTs displayed on their public profiles. Several boundaries have been identified: the positive impact of liking variety is diminished when buyers have better knowledge about the market (i.e., more trading experience) and about the creator (i.e., has previously bought the creators’ NFTs), or greater common follower overlap with the creator (i.e., more accounts follow both the buyer and the creator). These findings remain robust when addressing endogeneity using instrumental variables and when employing alternative measures based on deep learning–derived topic embeddings. Experimental evidence reveals the psychological process underlying the main effect that higher liking variety increases the perceived innovativeness, which enhances buyers’ evaluation of the NFT.
Temporary teams increasingly use GenAI to address innovation challenges under uncertainty and time pressure, but research explains more about what GenAI can do than how GenAI roles emerge as situational demands change. Drawing on interactionist role theory, we studied 28 temporary teams comprising 119 participants through stage-level interviews conducted during a 48-h innovation sprint. Our analysis identifies the servant, counselor, navigator, and oracle as role states that arise as teams use GenAI capabilities in response to shifting situational demands. We theorize this process as delegation calibration, whereby teams establish GenAI’s permitted influence, expected output, and intended reliance within a task episode. Calibration fit, that is, the degree to which the delegation aligns with task-episode conditions, shapes each role’s characteristic opportunity and challenge: conserved resources versus deviant realization (servant), expanded ingenuity versus trivial advice (counselor), harmonized insights versus eroded distinctiveness (navigator), and resolved deadlocks versus excessive complexity (oracle).
Generative AI shopping recommendation assistants (GSRAs) are reshaping online retailing, yet existing research largely treats GAI-based and search-based recommendations as independent alternatives. This study conceptualizes recommendation channel integration capacity (RCIC) and develops a framework explaining consumers’ intention to integrate heterogeneous recommendation channels. Drawing on service-dominant logic (SDL), affordance actualization (A-A) theory, and creative process theory (CPT), this study examines how perceived RCIC relates to perceived affordances and integrated recommendation usage intention. Using task-based survey data and PLS-SEM analysis, we find that perceived RCIC is positively associated with four perceived affordances—creativity, communication, explanation, and analysis—which subsequently strengthen consumer empowerment and value co-creation, increasing integrated recommendation usage intention. Necessary condition analysis further identifies consumer empowerment and value co-creation as necessary conditions for high levels of integration intention. By shifting the focus from technology substitution to recommendation channel integration, this study establishes an initial conceptual and nomological foundation for understanding the complementary use of emerging GAI-based and traditional search-based recommendations.
