By Ian C. Schick, PhD, JD, CEO & Co-founder of Specifio (originally posted here on Sept. 24, 2018)
Attorneys are increasingly adopting artificial intelligence to streamline the drafting process for various legal documents, but are they currently meeting the ethical guidelines for protecting private client information? Below, we discuss the role and ethical responsibility of attorneys using third party document automation services and recommended measures to protect confidential information.
The Model Rules of Professional Conduct 3.3 state that lawyers must make “reasonable efforts” to protect client information, which includes understanding the sensitivity of the information, potentiality of disclosure without employing additional measures, and cost of these measures. Lawyers are responsible for maintaining the confidentiality of the information to and from the document automation service. Hypertext Transfer Protocol Secure (HTTPS) and email with encrypted text or attachments are the two most secure methods for transferring information online.
The ethical obligations relating to lawyers outsourcing work to third party attorneys similarly applies to non-lawyers, such as machines. According to Model Rule 5.3, attorneys are responsible for ensuring that the ethical conduct of non-lawyers are identical to professional attorneys.
While reviewing written confidentiality agreements and security policies is recommended, the Model Rules make it clear it’s important lawyers understand what happens to confidential client data (1) when it is transmitted to a third party, (2) while the third party is in possession of the data, and (3) when the third party returns work product back to the lawyers.
Specifio is the first and only fully-automated patent drafting service. This means there is no human intervention, which addresses potential confidentiality and conflicts issues. In addition, all documents transmitted between the service and the user are password-encrypted.
In order to provide customer service and improve our technology, however, Specifio does keep metadata associated with the client information. The metadata includes a “content-stripped” version of the confidential information in which content words are replaced with nonspecific symbols, making the meaning of the text uninterpretable. These disguised content-stripped texts are only used to analyze document structure, word form, word count, and other metrics that are useful without knowing the underlying contents.
Below is an example of what content-stripped information may look like:
“The present disclosure relates to systems and methods for facilitating review of a confidential document by a non-privileged person by stripping away content and meaning from the document without human intervention such that only structural and/or grammatical information of the document are conveyed to the non-privileged person.”
“The p0018 d0017 r0019s to systems and methods for f0000ing r0001 of a c0002 d0003 by a n0004 p0005 by s0006ing a0007 c0008 and m0009 from the d0003 without h0010 i0011 such that only 20012 and/or g0013 i0014 of the d0003 are c0015ed to the n0004 p0005.”
Finally, with no training on its machine-learning models on confidential information, there is never “cross pollination” between patent applications we generated. In other words, text from one auto-generated patent application will never appear in another auto-generated patent application.
DISCLAIMER: The views and opinions expressed in this article are those of the author and (1) are not provided in the course of and do not create or constitute an attorney-client relationship, (2) are not intended as a solicitation, (3) are not intended to convey or constitute legal advice, and (4) are not a substitute for obtaining legal advice from a qualified attorney. You should not act upon any such information without first seeking qualified professional counsel on your specific matter. The hiring of an attorney is an important decision that should not be based solely upon Web site communications or advertisements.