The identification of privileged content remains one of the most time-intensive undertakings of any discovery project. In recent years, there have been significant developments in the ability to use analytics tools and artificial intelligence (AI) to identify the existence of privilege. This helps legal teams streamline the review process while maintaining privilege.

Below is an outline of steps to consider for handling the identification and review of potentially privileged content in antitrust matters. Utilizing advanced tools and following the appropriate steps can be a big time-saver when producing documents to the government during an antitrust enforcement action.

Step One: Compiling a Privilege Screen

The legal team (corporate and outside counsel) and the data provider should work closely to compile a “privilege screen” of attorney names and terms that can signify the existence of privileged data. The privilege search term report will show the total and unique number of “hits” for each name and term. Once the screen has been finalized, it is run against the data that is possibly responsive. As with any search term analysis, it is important to test the results of those search terms, particularly terms that could have more broad usage and may be generating false positives.

Step Two: Highlighting Terms Indicating Privilege

All “hits” are then highlighted in the review database. The team can also deploy custom analytics workflows to identify documents and document families that only have privilege terms appearing in the footers of emails, as it is often customary practice to include the “privileged and confidential” disclaimer on all emails.

Step Three: Deploying an AI Model

To reduce the number of false hits and expand the scope to include privileged documents the privileged terms might miss, counsel can also use analytics and AI tools to supplement the standard privilege screen. This starts with feeding sample coded documents into an AI system. These systems typically use the text and metadata of documents to score each on a scale from 0 to 100 – the higher the score, the more likely the document is privileged. The scores provide another means of identifying privileged documents and help to reduce the number of false hits returned by the standard privilege screen.

A privilege expert (typically from the review provider) can then train the system using examples from the collection. The training effort is typically less than 3,000 documents and continues while the system is gaining reasonable value from additional training. The expert also uses social network analysis, domain analysis, and other analytics tools to identify potential privilege actors. Statistical sampling and targeted searches of the null set are then used to validate the results of the privilege AI model.

Step Four: Sorting the Documents

After deploying the AI model, documents identified as possibly privileged are sorted into two groups:

1. Higher Probability Documents are those containing certain attributes that typically signal the existence of privileged content. Common examples include communications with outside legal counsel for the end client and communications with internal counsel. These documents can be moved directly to a privilege log review and redaction workflow, which obviates the need for first-level privilege review. The reviewer conducting the privilege log review will then confirm the existence of privilege.

2. Lower Probability Documents are those that have some indicia of privileged content, but still require validation through the first-level review process to confirm the existence of privilege. Outside counsel will decide whether to review all of the documents flagged as potentially privileged or to cut off review for those with lower probability scores.

Step Five: Identifying Additional Privilege Sources

Once the review starts, the team will identify additional names of individuals or organizations that may create or break privilege. New privilege names are escalated to outside counsel to determine whether they need to be added to a privilege search term report. Any new names added to the privilege search term report will then be normalized across previously reviewed, non-privileged documents.

Conclusion

The combination of traditional and rigorous screening with AI privilege tools will provide the best results when reviewing for privilege in antitrust matters. Additionally, the process involved is defensible. Counsel can easily explain the process above to a regulator who questions methodology. Coupled with the protection of a protective order regarding the inadvertent production of privileged material, this provides a thorough workflow for identifying and properly logging privileged data.

This blog post is derived from the Chapter titled “Outsourced Document Review:  Data Intelligence, Technologist Lawyers, Advocacy Support” by Edward Burke and Allison Dunham, which appears in the Thomson Reuters treatise eDiscovery for Corporate Counsel (2022). Reprinted with permission, © 2022, Thomson Reuters. Jason Butler also contributed to this blog.

A link to the book appears below:

https://store.legal.thomsonreuters.com/law-products/Treatises/eDiscovery-for-Corporate-Counsel-2022-ed/p/106770308