Blog

Making the Case for Good Old Fashioned AI in E-Discovery

Generative AI is in the news, but there are still plenty of ways to get value from older types of AI like machine learning.

Artificial intelligence (AI) is dominating headlines today, especially with the rise of generative AI tools like ChatGPT, Google Bard, DALL·E, and MidJourney. These technologies have sparked everything from excitement about productivity gains to concerns about risks and misuse. However, AI itself is not new—it has evolved through multiple stages, each with distinct capabilities and applications.

The Evolution of AI

1. Rule-Based AI

The earliest form of AI relied on hard-coded rules created by programmers.

Systems followed predefined logic to perform tasks
Effective for structured problems like games, speech recognition, and text analysis
Example: IBM’s Deep Blue, which defeated a world chess champion

While powerful in narrow domains, rule-based systems lacked adaptability.

2. Machine Learning & Deep Learning

The next wave introduced machine learning (ML)—systems that learn from data rather than explicit rules.

Uses algorithms to identify patterns and make predictions
Improves over time with more data
Applications include spam filtering, recommendation engines, and e-discovery

Deep learning, a more advanced form of ML, leverages large datasets and computing power to deliver more accurate and sophisticated results.

In legal and compliance contexts, these technologies power tools like:

Predictive coding
Technology-Assisted Review (TAR and TAR 2.0)

These tools significantly reduce time and cost in document review processes.

3. Generative AI

Generative AI is a subset of deep learning that creates new content based on patterns learned from training data.

Generates text, images, and other media
Useful for drafting documents, summarizing information, and creative tasks
Widely used but still prone to errors (e.g., fabricated citations in legal filings)

AI in Legal and E-Discovery Workflows

While generative AI gets most of the attention, proven AI technologies like machine learning and TAR remain highly valuable—especially when used alongside human expertise rather than replacing it.

Organizations can safely leverage AI by applying it to:

Prioritizing documents for human review
Conducting privilege reviews
Reviewing incoming document productions
Performing secondary reviews on lower-priority materials
Running quality control checks on human reviewers

This hybrid approach balances efficiency with accuracy and reduces risk.

Key Takeaway

AI is not a single technology but a continuum—from rule-based systems to machine learning to generative AI. While newer tools attract attention, established AI methods already deliver measurable value, particularly in legal and e-discovery contexts.

Whether organizations fully embrace AI or take a cautious approach, its ability to reduce time and cost while improving efficiency makes it increasingly difficult to ignore.

‍