Who Owns the AI Lawyer?Copyright and Authorship in AI-Generated Legal Content

Author~ Shrishail Lakhepatil

Place yourself in a scenario where you hand over the process of drafting the Non-Disclosure Agreement to the AI legal platform. Within seconds you will have a highly professional document, replete with all kinds of clauses that you can desire. You take it, sign and forward it. It works like a charm. But there remains an important question hovering over both you and the platform alike: whose creation is this?

It should come as no surprise if I say that this is not a moot point at all, considering the times we live in. All kinds of AI software have entered the market from litigation research to contract drafting purposes. This is more about regulation of technologies and the future of legal services than anything else. Intellectual Property Law still remains clueless on this front.

I. The Rise of the AI Lawyer

The growth of legal technology in India has been rapid. There are legal tech tools today that go beyond aiding in drafting a legal brief or assisting in litigation. With the power of artificial intelligence, they create, they generate contracts, compliance reports, analyses of judicial opinions and predictions on a verdict. The result is a structured document – a brief or a memo or policy report that you will not be able to distinguish from one drafted manually by humans.

What this leads to is a rather tricky situation when it comes to IP law. In case an AI tool creates the brief for you, is it protected under copyright? If so, to whom does the copyright belong to?

II. What the Copyright Act, 1957 Actually Says

The rules pertaining to the rights of original works, according to India, have been established under the Copyright Act, 1957. This act, under section 13, makes it explicitly clear that protection can be granted to ‘original literary works’, which is inclusive of things like opinions, agreements, and legal briefs.

Yet, originality must be demonstrated. According to the Supreme Court ruling, Eastern Book Company v. D.B. Modak, in the year 2008, it must be the result of the author’s intellectual effort and should possess a level of creativity. In other words, it should not be something produced as the result of mere labour, as opposed to the use of skills and judgement. However, an AI does not possess skills and judgement to begin with. Hence, it cannot provide the same.

Section 2(d), too, proves to be of critical importance here. This part of the act clearly specifies that ‘author’ refers to a human being. Therefore, no AI can be considered one, thereby disqualifying it from being the author of any creative content.

III. The Three-Way Ownership Problem

Assuming that the output created by AI tools falls under copyright protection, the next question we need to ask is: Who owns it? It is possible to justify copyright ownership on behalf of three different parties.

1. The User or Client

The lawyer, litigant or corporate entity who provided the input will argue that the resulting work is theirs as they were in charge of the entire process. Their position has merit since the user was responsible for the terms of input, selecting the tool as well as framing the question. However, copyright is linked with authorship and without sufficient creative contribution by the user, the notion of authorship simply doesn’t hold any water.

2. The AI Developer / Platform

Alternatively, there’s the firm that developed the platform. Since they employed the scientists to build such a platform, and they have spent extensive time and resources into training the algorithm and customizing the architecture, they might be justified in claiming copyright ownership. However, copyrights protect expressions – not effort. The former idea was effectively debunked in the Eastern Book Company vs D.B Modi & Ors case, where the Indian Supreme Court found sweat of the brow doctrine untenable.

3. The Public Domain

In case both fail the criteria of authorship, the resultant document will merely fall into the category of the public domain works. This is definitely not good news for the law firm who would have paid for an artificial intelligence generated contract. Anyone would now be free to replicate or modify the document without having to worry about the consequences.

IV. The Gap Created by Computer-Generated Works

The gap created by computer-generated works in countries like India has been addressed through laws and legislation. The Copyrights, Designs and Patents Act 1988 of the United Kingdom is an example of this kind of legislation. According to section 9(3), the copyright in computer-generated work will vest in the person who arranges for its creation. Normally, this would be the developer or operator of the computer. It is a convenient way of resolving disputes regarding the creator.

Unfortunately, India does not have anything like this. Its Copyright Act 1957 is old and predates generative AI technology and thus fails to make provision for the copyright of computer-generated works. The only document one can refer to is the National IPR Policy published in 2016 or the 2021 Report of the Parliamentary Standing Committee on Commerce.

V. Practical Implications for Legal Practice

The fact remains that this will be no academic undertaking. The issues at stake are quite real and the results very immediate.

Take, for instance, the problem of confidentiality and data privacy. The client presents his facts to a lawyer who uses them as inputs to an AI to receive a legal opinion. Who, then, owns the resulting document? If the platform owns it, doesn’t the lawyer owe confidentiality to the client and prevent the use of the data by the platform for training their next model? Also, consider the Digital Personal Data Protection Act of 2023 and all the obligations on data fiduciaries for handling data provided by users to an AI system.

Professional ethics may come under scrutiny too. The Bar Council of India mandates that the lawyer be skilled and act in the best interests of the client. However, if there is a flaw in a document prepared using an AI tool, and copyright fails to protect against it, can the client claim any ownership?

For the time being, most of these platforms will take an approach through contracts since the Copyright Act provides no clarity regarding the issue. The Terms of Service of these platforms would typically provide for the transfer of copyright in the output generated to the user while retaining copyrights in the underlying model with the developer. However, you can assign something only if there is something to assign – if the Copyright Act does not acknowledge anyone as the author of the work, then no contract can assign its copyright.

VI. The Reform

India needs some legislative candor on account of three issues.

First, there is a need to amend the Copyright Act in order to treat 'computer-generated works' as distinct from human creations and to hold that the copyright in computer-generated works would vest in the person making the arrangements for creation of these works, similar to how it happens in the UK. This removes any need for debates on consciousness of AI and brings certainty in the market.

Second, we might consider putting in place a graded period of protection in such cases. Why grant life-plus-sixty years in a computer-generated work?

But there must also be an obligation to state where AI has been involved in preparing the documents or pleadings. It is only fair that regulators and clients be told if what they see is the lawyer’s judgment or the output of an algorithm. The objective must be clarity, and not a blanket prohibition.

VII. Conclusion

In India, the AI lawyer is no longer fiction. She advises on risks and churns out compliance reports and draft agreements. But what Indian law has still to come to grips with is who owns the works produced by such technology.

By the standards of the 1957 Copyright Act, any document prepared by AI in India remains in the shadow zone, devoid of any legal status and ownership for practical purposes. In a field of fine lines, it is quite disconcerting.

The face of legal practice has already undergone transformation due to AI. The bigger challenge ahead lies in how Indian intellectual property law will adapt to meet the challenge, or if we shall leave this work of AI out in the cold at the doors of the courthouse.

Note - The information contained in this blog is for general informational purposes only. We endeavour to keep all content accurate, updated, and free from any form of misinformation or objectionable material. However, we shall not be responsible for any claims arising out of copyright infringement, plagiarism, or related issues; such responsibility lies solely with the respective authors. If you find any misinformation or objectionable content on this website, please report it to us at: editors.ilw@gmail.com