Four Areas Companies Must Focus On For Artificial Intelligence Auditing

A new research report published by the MIT Center for Information Systems Research (MIT CISR) has identified four key areas where companies need to ensure auditing and visibility in artificial intelligence projects. Based on interviews with 48 leaders from 28 organizations worldwide, the study highlighted architectural compliance, innovation management, workforce readiness, and vendor management.

While artificial intelligence technologies make it easier for organizations to experiment, prototype, and develop new applications, they can also lead to unnecessary investments, non-scalable applications, and unexpected costs. Since establishing separate auditing mechanisms for every artificial intelligence application is not sustainable, companies are advised to focus on areas where auditing will create the greatest impact. Researchers Alan Thorogood, Peter Reynolds, and Asif Gill listed four main areas where companies can maintain sufficient control without slowing down their artificial intelligence efforts:

Architectural compliance accelerates business units in developing and deploying applications through a strong artificial intelligence architecture while establishing clearer control points for the organization. In the area of innovation management, it is emphasized that decentralized experiments can lead to overlapping projects or prototypes that are too expensive to scale, and therefore this process needs to be managed.

For workforce talent, it is stated that merely hiring specialists is not enough; existing employees must be retrained, and roles must be redesigned for jobs that will be augmented by artificial intelligence. In the context of vendor management, as companies become more dependent on external suppliers, new questions are reported to arise regarding third-party risks, changing capabilities, and costs.