Executive Summary

The biggest challenge to the broader adoption of scalable AI technologies is not better algorithms and models but instead getting access to more high-quality training data. In this white paper, we discuss the challenges inherent to obtaining high- quality data and our view of how this field is evolving in light of advances in new technologies and more stringent data security and privacy regulations. In addition to high-quality, security, and privacy, we emphasize the need for scalability, speed, and flexibility to meet the requirements for larger and larger volumes of data to train AI algorithms. After briefly presenting current effort to deal with these requirements, we show, through examples, how technology and machine learning can assist in data annotation and provide outstanding results when focusing on the combination of what humans and machines do best.

Specifically, machine learning models can improve data quality and data throughput, while a data annotation platform integrating various technologies can scale with customer requirements. Via customization, a welldesigned platform can help find the sweet spot between quality, cost, and speed. Considering future requirements for scalability and flexibility, we conclude that the best annotation company is the one that can move quickly from one of the extremes of the continuum (automatic annotation) to the other (manual annotation with high-quality annotators) to satisfy customer requirements. Data annotation will serve AI engines’ future needs by optimally combining technology and human expertise and refining, as technology progresses, this combination.