Peer to Peer Magazine

Summer 2019: Part 2

The quarterly publication of the International Legal Technology Association

Issue link:

Contents of this Issue


Page 50 of 63

52 F or better or for worse, eDiscovery languishes far from the forefront of artificial intelligence (AI) applications and development. To be sure, AI already plays a key role in driving efficiencies in the eDiscovery process. But the adoption and implementation of advanced artificial intelligence techniques, that indeed are intended to substitute for direct human observation and analysis in the eDiscovery process, continues to proceed at what sometimes seems to be a glacial pace. This paper takes a critical look at two advanced artificial intelligence applications – natural language processing (NLP) and image recognition – and provides a pragmatic assessment of their current and eventual integration into the eDiscovery process. Where are we today? With all the hype surrounding the potential for artificial intelligence in eDiscovery, it is easy to lose sight of the fact that AI has been driving eDiscovery efficiencies for quite some time now. Most modern technolo- assisted review techniques rely on supervised machine learning. Earlier document and feature aggregation techniques like clustering and latent semantic indexing relied (and continue to rely) primarily on unsupervised machine learning. And both machine learning approaches are indeed AI systems. Nevertheless, advanced AI applications like NLP and image recognition remain in their nascent stage. Both provide a measure of superficial utility in eDiscovery, but their true potential has yet to be fully recognized and exploited. Notwithstanding the level of excitement, progress toward more fully integrating NLP and image recognition is likely to be, and for all intents and purposes probably should be, incremental and thoroughly considered. Artificial intelligence applications should only be incorporated into the eDiscovery process if and when they can provide a demonstrable benefit to the objectives of the process, i.e., finding the documents necessary The Future of AI in eDiscovery B Y T H O M A S C . G R I C K S I I I , A N D R E W B Y E , A N D J E R E M Y P I C K E N S , P H . D . In collaboration with the ILTA Litigation Support Content Coordinating Team

Articles in this issue

Archives of this issue

view archives of Peer to Peer Magazine - Summer 2019: Part 2