Generative AI for E-Discovery: Adoption and Trends in Germany

Currently, leading use cases include:

• An AI quality control sweep of human review outputs to quickly identify conflicts between the reviewers’ categorisation and the AI’s.
• Analysis of third-party disclosures.
• Reviewing highly technical documentation (patent and pharmaceutical cases) where human reviewers face challenges in parsing information out of complex or nuanced documents.
• Automated PII identification and redaction workflows, particularly valuable for cross-border data transfers which can require extensive manual redaction.
• Chronology generation for key logs or interview preparation.

FTI Technology recently hosted a generative AI training event in Berlin, where experts presented practical applications of generative AI-powered legal workflows through a simulated investigation of a failed construction project. Using IQ.AI by FTI Technology™, the trainers demonstrated how generative AI can efficiently classify documents for relevance and automatically generate preliminary summaries and chronologies from the relevant materials.

The dataset was designed to contain documents that would require careful deliberation by a human reviewer to make a relevance designation. These nuanced documents illustrated the effort and strategy required to draft a prompt that would identify a similar or higher proportion of relevant documents than a human reviewer. 

Event guests shared their experiences and emphasized that practical adoption relies on the seamless integration of generative AI technology within the tools that they already use. As noted by Omair Bajwa, BioNTech Manager, Compliance & Business Ethics, "Access to these tools must exist at the document level, with security and privacy built in by default." This sentiment echoed throughout the discussions — with the strongest demand existing for solutions that provide:

1. Direct integration with existing review platforms.
2. Germany and European Union-based hosting with robust data privacy guarantees.

Additionally, legal professionals in Germany already have a mature understanding of the capabilities of generative AI and the majority particularly value expertise in designing and implementing user-led, cost-effective workflows that enhance accuracy and efficiency by leveraging both new and existing tools. For example, one participating lawyer illustrated this point, noting, “We understand the complexities and use cases for using generative AI to accelerate a first level review and have seen fantastic recall when the generative AI models are prompted correctly. However, a high number of false positives is still to be expected, and it’s therefore important to work closely with your legal technology team to anticipate and find ways to deal with this.”

At each event FTI Technology experts attend and host, there is an evident growing interest in high-value use cases where generative AI can be used to drive efficiencies in legal review.  Recent experience across client engagements here in the country and broadly demonstrates a strong preference in targeted applications that address specific pain points. Particularly, those combining efficiency gains in running quality control sweeps, creating key logs and chronologies, extracting data and identifying personally identifiable information. Rather than displacing current processes, successful adoption will involve hybrid models that leverage AI's strengths while mitigating its limitations through human oversight and traditional and established legal techniques.

Related topics:

The views expressed herein are those of the author(s) and not necessarily the views of FTI Consulting, its management, its subsidiaries, its affiliates, or its other professionals.