MIT CISR’s data research team has been tracking the journeys of fifty-two AI projects since 2019. These journeys have helped the team identify two distinct patterns of AI model growth: scale up (i.e., increase in volume of core model use) and scale out (i.e., increase in number of recontextualized models). We have observed, however, that many of the AI initiatives are not scaling as desired. Likely, there are new managerial practices that leaders need to draw on to advance each AI scale dimension more effectively. In 2022, we will examine AI projects that report scale success and create practical insights to inform successful AI scaling. We will ask questions such as:
- How do companies successfully manage the processes of AI scale up and scale out?
- What is the impact of scaling AI on data-driven value creation and capture
Methods: Interviews with a subset of the fifty-two AI projects from our 2019 sample and a case study or vignette of one or two AI projects that demonstrate effective AI scaling practices
Seeking: Companies that believe they are successfully scaling AI—both up and out
Contact: Ida A. Someh
