[author: Doug Austin, Editor of eDiscovery Today]
Even though we’re far from achieving critical mass in the legal profession when it comes to the use of predictive coding technologies and approaches in electronic discovery, the use of predictive coding for document review – especially relevancy review – to support discovery is certainly the most common use of artificial intelligence (AI) and machine learning technologies. Some of you reading this blog post may be “old pros” at this point when it comes to the use of predictive coding while others of you still have yet to “dip your toes” into the predictive coding pool.
But applying machine learning technology to support document review (which is predictive coding) is far from the only discovery-related workflow and use case where AI and machine learning technology can be applied. There are several others that forward-thinking organizations are looking to also implement to streamline workflows in the discovery life cycle.
With that in mind, here are three use cases of AI and machine learning technology you may not know:
Information Governance and Defensible Deletion
How could we forget one of the “forgotten ends” that I discussed last week?
An effective Information Governance program has become a “must have” to achieving effective discovery downstream. Why? Because of the Big Data challenge (that I also discussed here last year).
When the amount of data in the world grows 1,630 times over 20 years, that tells you all you need to know about why InfoGov has become such an important part of not just the discovery lifecycle, but also the management of information for all organization activities. Data is simply overwhelming without a program to effectively manage it – including removing the information your organization doesn’t need on a timely basis.
AI and machine learning technologies have become key to helping organizations understand their data sooner. More and more, organizations are implementing AI technologies and classifiers to support records management and defensible disposal initiatives within the organization.
For example, identification of redundant, obsolete, or trivial (ROT) data (which can be defensibly deleted) within your organization can be automated by AI technologies and extend human identification of ROT data to other ROT data that doesn’t have to be identified by humans.
As it’s more important than ever to identify ROT data to eliminate the risk of breaches, the downstream benefits are huge for using AI in this way.
Investigations and Sentiment Analysis
Corporate investigations are on the rise and the reasons for them expanded even more during the pandemic. For example, in a September 2020 survey conducted by the Association of Certified Fraud Examiners (ACFE), 74% of surveyed certified fraud examiners indicated that preventing, detecting, and investigating fraud in the COVID-19 era had become even more challenging than it was before the pandemic.
Sentiment analysis involves finding key words and phrases relating to sentiments or feelings that indicate potential detection of employees engaged in fraud or harassment (for example, words that indicate anger or frustration). AI and machine learning technologies help automate the identification of communications more likely to indicate activities that may require investigation.
Data Privacy and Identification of PII
Unless you’ve been living under a rock, you’re aware that the data privacy landscape within the world continues to change. In the past few years, we’ve seen the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA) go into effect, among other data privacy laws, and plenty of other laws have at least been enacted around the country and the world. There is not only more information to manage within organizations, the expectations for protecting sensitive data have also expanded for all organizations, whether or not they have litigation.
As a result, the identification of personally identifiable information (PII) has become important in a variety of situations. Organizations are even having to implement brand new workflows to respond to Data Subject Access Requests (DSARs) from individuals where they request information about the way companies handle their personal data. The ability to identify PII within an organization has become paramount to support these requests and other data privacy requirements.
Once again, the application of AI and machine learning technologies can help here.
Social security numbers, phone numbers, driver’s license numbers, and credit card numbers are easier to identify through pattern matching, but names, addresses, and health information are much more difficult to differentiate.
Take my last name for example – Austin–which also happens to be a city in Texas. AI can help differentiate the true matches to data related to me based on context from the false ones that are just related to this capital city.
Identification of PII is an iterative process that leverages AI technology to identify that PII much more quickly throughout an organization’s data corpus to address ever changing data privacy requirements.
Leveraging AI and machine learning technologies in legal and eDiscovery is about much more than predictive coding today – it’s about more use cases than ever as legal professionals learn how to fully utilize (and harness) the technology properly.
The AI/machine learning train for legal and eDiscovery use cases is building up steam – are you on it?