This study employs path dependence theory to study the impacts on changes in newsroom practices and strategic behavior of newsroom managers. We use Sydow et al.’s (2012) organizational path dependence model and define AI innovation as (1) a dynamic process that is shaped by multiple factors of innovation; (2) causes path dependence when triggered by historical events leading to a critical juncture; (3) governed by a regime of positive, self-reinforcing feedback constituting a specific pattern of social practices, which gains more and more predominance against alternatives; and (4) leads, at least potentially, into an organizational lock-in, understood as a corridor of limited scope of action that is strategically inefficient; and (5) eventually leads to strategic action to break away from a chosen path and take strategic action to avoid any dismal dependence and thus inefficiency. The empirical analysis follows a qualitative case study research design to account for emergent historical insights into path dependence in AI innovation gained over 3 years. Two different groups of participants were interviewed about historical factors shaping and constraining AI innovation: (1) news media executives (n = 10), including Austria’s public service broadcasting corporation ORF, the country’s national news media agency APA, managers of regional high-quality newspapers, and digital AI news media innovators; and (2) academic experts in the field (n = 11). Managerial perceptions were evaluated alongside four phases of path constitution practices (preformation, path creation, lock-in, and path release) and three path dependence mechanisms (historical events, critical juncture, and strategic action). We found the evolution and constitution of an AI innovation path in news media to be an ambiguous concept and an overambitious practice. While AI raises opportunities for competitive advantage, it also leverages the complexity of challenges faced by journalists and newsroom executives. The analysis further proved that executives preferred the constitution of an idiosyncratic path and proposed AI innovation as a mindful strategy for breaking legacy paths. History mattered, but the actors also shaped it.