• by Joe Mandese , Yesterday

The Media Rating Council, which is in the process of a major reevaluation of its existing standards and guidelines related to the acceleration of artificial intelligence (AI) and machine-learning models and their impact on media measurement, this afternoon released some interim guidance focused on nine core principles related to AI in media measurement:

• Explainability: The models and algorithms outputs are explainable and allow for validation.

• Compliance: The model and algorithms used comply with applicable MRC Standards and Guidelines.

• Fairness: The potential model bias is assessed for group and individual fairness in representation and model outputs.

• Transparency: The purpose and design of the model are disclosed to users in sufficient detail to understand the outputs produced by the model.

• Security: The systems and algorithms involved in operating the model, including data transfers and storage, are protected against unauthorized access.

• Data Protection: The system’s data use aligns with permitted rights and confidentiality.

• Reliability: The models are capable of producing valid and accurate outputs based on stated expectations for the model.

• Accountability: The model inputs, algorithmic designs, processes, thresholds, limitations and assumptions are known to the measurement services and quantified, as applicable.

• Adaptability: The model is adaptable to technological advancements and changes in media consumption, content delivery and consumer behaviors.

Details of the MRC's guidance related to the principles, as well as existing media-measurement methods, can be read here, but the MRC identified a half dozen "priorities" for where "incremental guidance" and requirements are needed, because they are not covered by the MRC's existing guidelines and standards:

• General

• Invalid Traffic

• Brand Safety

• Base Digital Measurement

• Identity and Big Data

• Auctions

The MRC said it already has engaged with audit firms that are actively conducting MRC audits to "build a list of standards and guidelines currently in use to evaluate AI use in various different areas," and published the following table itemizing where its AI principles impact existing guidelines and standards: