The Technical Committee for Standardisation of Artificial Intelligence ('TK 164') of the National Centre for Digital Economy of Moscow State University requested, on 20 September 2022, public comments on draft standards in the field of data quality for artificial intelligence ('AI') and machine learning. In particular, the TK 164 highlighted that five standards on AI and machine learning requirements have been drafted for public discussion. Furthermore, the TK 164 outlined that the five standards include a draft standard on the structure of the data life cycle, and a draft standard regarding the structure of data lifecycle, split into four parts on:

• overview, terminology and examples;
• indicators of data quality;
• requirements and guidelines for data quality management; and
• tools for monitoring data quality.

More specifically, the standard on indicators of data quality establishes the characteristics of data quality, including, among other things, portability, relevance, and identification of data. Further, the same standard details techniques for improving data quality. On the other hand, the standard on requirements and guidelines for data quality management extensively details recommendations for the whole data lifecycle, including data acquisition and pre-processing, alongside requirements regarding the purpose and intended use of data collected. Finally, the standard on tools for monitoring data quality also provides recommendations on the outsourcing of AI data and use of cloud services for such data. Likewise, other tools for monitoring data quality are provided to include, among others, clearing data, filtering data, and data anonymisation.

Public comments may be submitted to [email protected] until 14 November 2022.

You can read the draft standards, only available in Russian, here.