I. INTRODUCTION
The winds of change are blowing all over. In today’s fast-paced world, wherein we see new technological advancements coming in each day, it is difficult for us to even imagine our lives without technology. It cannot be denied that the biggest technological advancement today is artificial intelligence (hereinafter also referred to as “AI”). We all must have watched sci-fi movies depicting robots and synthetic intelligence, but the question for us is: What if such super advanced technology that we witness in such movies come to reality? What would be its practical implications with regard to our lives? And most importantly, how would the rights and liabilities concerning AI be deciphered? The country wherein courts are bombarded with a high pendency of cases including those of rapes and murders, it is highly crucial to reflect: Are we really equipped with an efficient strategy and legal framework so as to deal with the multifarious aspects of AI? This article attempts to dive deep into various facets of AI and to analyse the same in light of the legal paradigm. The legal issues concerning artificial intelligence that would be discussed hereinafter would compel us to rewire our thought process and create a pathway for us to create a holistic framework to deal with the upcoming technological concerns.
II. GENERATIVE AI SIMPLIFIED
In ordinary parlance, artificial intelligence can be defined as the cognitive intelligence of a machine. Let’s try to understand this with an example. ‘Siri’ and ‘Alexa’, commonly found in households nowadays, are prominent artificial intelligence programs. ‘ChatGPT,’ a prominent artificial intelligence chatbot, launched on November 30, 2022 by an artificial intelligence research company named OpenAI, is designed to converse with humans in a conversational way, thereby generating textual content such as essays, articles, posts for social media, emails, etc. It is pertinent to mention that ChatGPT is essentially a form of generative AI. Generative AI uses prompts fed into it by external sources, such as humans, and processes information to generate images, text, or videos. GPT in ChatGPT refers to ‘Generative Pre-Trained Transformer’. Based on an extensive algorithm or ‘training’ by human trainers, ChatGPT successfully generates texts and responses similar to those of humans.[i]
When it comes to the legal sphere, the significance of AI cannot be undermined. As an advocate, in order to file a petition in the court, with annexures having 500 pages, a hefty amount of work is required to be performed. It cannot be denied that reading those 500 pages would be a time taking exercise. But what happens if you get a gadget or a software which reads those 500 pages and gives you a summary of it? What if that machine takes up only 10 minutes to give you its salient analysis while deciphering its key areas? This is the essence of AI. It efficiently processes the information and gives you the desired output within a few minutes using its cognitive intelligence and analysing capabilities.[ii]
III. USAGE OF AI VIS-A-VIS LEGAL PROFESSION: GENERATION OF FAKE CASES
Imagine an advocate getting a knock on his door at 9 PM, and being apprised that an urgent lawsuit is to be filed the next morning in the court. Now it is only a few hours left in his hands to prepare the entire lawsuit from scratch, with a negligible amount of time to conduct comprehensive legal research work in order to substantiate his pleadings. Now, what if he asks an AI program to cull out relevant precedents pertaining to the facts of the matter that he entered into the program, and the program elucidates various case laws settled on that subject matter within seconds? So, this advocate does the same and incorporates those precedents in his pleadings to substantiate his lawsuit, and files it the next morning, as instructed. After a few days, he is apprised of the fact that the precedents he incorporated in his pleading are entirely fabricated and do not exist in reality, meaning thereby that the same was concocted by the AI program. Now what must this advocate do? He may either plead to amend his pleading with the leave of the court or he may plead to withdraw the lawsuit, prepare a fresh and verified one and then file it. However, it cannot be denied that in both the cases, not only has the precious time of the court been wasted, but the advocate too had to go through the pangs of mortification which could have been easily avoided with a little diligent research.[iii]
Similar to this happened in an American case involving an attorney named Mr. Steven Schwartz, who’s associated with a reputed law firm which is into civil aviation litigation. Allegedly, he was supposed to file a written submission before an American district court in an aviation injury claim matter and because of paucity of time, he relied on an AI program: ChatGPT for his legal research work, including identification of relevant settled cases. The program, having been fed in the proposition, formulated the written submission in a few seconds, quoting several decisions propounded by the US courts. The document so drafted by the program is filed before the federal court. During the arguments the opposite counsel apprised the court of the fact that he was unable to find the judgements quoted by the firm and requested the court to ask them to provide copies of the same. The court eventually found those judgements to be fabricated and that neither the parties so mentioned nor the factual matrices existed in reality.[iv] For this, the attorney had to issue an apology before the court, stating “I suffered both professionally and personally [because of] the widespread publicity this issue has generated. I am both embarrassed, humiliated and extremely remorseful.” [v]
Following this episode, the US courts highlighted its concern regarding the usage of these AI programmes by legal professionals. They pointed out that a basic amount of due diligence is required to be performed by attorneys and that they must verify the authenticity of the output generated by these programs before relying upon it completely. In the present matter, the court imposed a $5,000/- fine on Mr. Schwartz, Mr. Peter LoDuca and the law firm for relying upon fictitious research work.[vi]
This situation poses a great challenge before the legal fraternity, where on one hand, legal professionals are enthused with the advent of AI, engaged in predicting its miraculous uses in their legal practice, on the other hand we see cases wherein fictitious and fabricated out is getting generated by AI, attracting courts’ attention. The legal fraternity is therefore facing a dilemma with regard to the use of AI in the legal realm.
Interestingly, in order to tackle such issues, the US courts have asked for certification from attorneys to the effect that their submissions are reliable. Attorneys are supposed to certify that their work is not a generation of AI, or even if it is, they have verified and authenticated the same diligently. This preventive measure follows numerous episodes of generation of fake cases by AI. So, the lawyers would have to certify under their signature that whatever submissions they are filing have been duly verified.[vii]
IV. AI & IPR INFRINGEMENT
One of the prominent challenges concerning artificial intelligence circumscribes matters relating to the infringement of intellectual property rights.
For any ‘work’ generated by AI, the prime issue stems from the point pertaining to ‘originality’ of the work. Further, who must be entitled to copyright also reflects a scope for discussion. Would it be the individual entering prompts into the program and receiving an output therefrom? Or would it be the concerned AI program which in the first place generated such work using its cognitive intelligence? In light of the above mentioned, it all comes down to two basic facets: (1) Entitlement of Copyright, and (2) Liable in case of infringement.
Another challenge involving copyright is with regard to the ‘output’ given by the generative artificial intelligence and its compliance with IPR norms. For instance, it is difficult to ascertain cent percent whether the output given by AI is an original work or a mere copy of someone else’s original literary work. It essentially deals with the training mechanism of the artificial intelligence program, i.e., what content is being fed into its algorithm to train the program for generating output, based on prompts fed by individuals.
It is pertinent to mention here that companies in the USA have filed lawsuits against these giant AI Tech companies, alleging the use of their copyrighted work for the purpose of training these artificial intelligence chatbots. Several authors have raised their voice against these companies stating that their original literary works have been used to train their AI algorithm without taking their prior consent or approval. Recently, The Times sued OpenAI and Microsoft for alleged copyright infringement, stating that no approval was granted for using their work to train AI programs.[viii] Amidst the pendency of such suits, it would be worthwhile to observe subsequent developments in the AI vis-a-vis IP regime as most of it has been left to courts to decide and adjudicate upon.
Furthermore, another aspect is regarding the liability of the service provider. In the USA, would it be sufficient for a service provider to seek exemption from liability and accountability on the ground that he is governed by section 230 of the Communications Decency Act, 1996? Or, in the Indian context, would the artificial intelligence coder be able to take benefit of section 79 of the Information Technology Act, 2000 stating that he is a mere intermediary, thereby incurring no liability? Ultimately, the judicial system needs to play a proactive role in interpreting the existing laws to deal with the present-day scenario.[ix]
V. CAN AI BE TREATED AS AN AUTHOR OR INVENTOR?
The fundamental conundrum when it comes to ‘authorship’ of a work concerned with an output generated by an AI program essentially revolves around a basic question: If an essay has been generated by ChatGPT, who would be the author of that work? Would it be ChatGPT itself who originally generated the work using its algorithm and processing abilities, or would it be that individual who, using his intellect, fed in tailored prompts so as to get his desired output from the program?
Recently, a sudden boom was observed in the AI written e-books on Amazon. Essentially, these books were generated by AI programs and were published under the name of individual persons on Amazon and the authors thereof acknowledged and recognised ChatGPT and the like chatbots, in creation of their works.[x] However, if in future, any inadequacy with regard to copyright compliance is found therein, then the fundamental problem would arise as to who should be made liable for copyright infringement. Would it be the liability of the individual person or the AI program? Although, the authors must exercise due diligence in checking the veracity of the output generated by the AI programs, it can nonetheless be worthy to ask as to who shall be made accountable in case of an infringement or deficiency subsequently found therein.
Likewise, when we ponder upon the concept of ‘inventions’, there are three basic requirements which need to be fulfilled to obtain a patent: (1) Novelty, (2) Inventive step, and (3) Industrial application. Now the fundamental question being: Whether an AI can be granted the status of being an ‘inventor’ if all these three requirements are being fulfilled? After scrutinising the provisions of the Patents Act, 1970, one can understand that a machine cannot be granted status of an inventor as per Section 2(p) thereof, which envisages a ‘person’ to be a patentee, with no room left for an AI algorithm to fit into this definition.[xi]
VI. IS THERE ANY LAW ON AI?
There is no law internationally to govern issues raised by artificial intelligence yet. However, the United Nations Ad Hoc Committee on Cybercrime is proposing the draft of an international convention governing the potential misuse of communication technologies for criminal purposes which could possibly cover a few key areas concerning the use of artificial intelligence.[xii] On the other hand, China has already brought forth the first ever law on generative artificial intelligence and the same was implemented on August 15, 2023. This legislation is a comprehensive one, dealing with various aspects concerning the use of AI including the need for auditing, i.e. submission of an AI program for approval of State authorities before being released for public use, various rights, and liabilities of stakeholders, etc. A few months down the line, the European Union in December 2023 passed the law on artificial intelligence with special emphasis on mitigating public harm caused by such technology by way of encouraging accountability and transparency.[xiii]Furthermore, the Illinois Artificial Intelligence Video Interview Act came into effect in 2020 governing employers who use AI tools to evaluate the video interviews for various positions based in Illinois.[xiv]
India presently does not have a dedicated law on artificial intelligence, however it does have a national policy of 2018 named ‘National Strategy for Artificial Intelligence’ for adoption, reskilling and research on AI.[xv] However at the moment, the Information Technology Act, 2000 could possibly cover a few aspects surrounding artificial intelligence due to the fact that an AI algorithm is more or less a ‘computer program’ which is in turn governed by the said act. Having mentioned that, it cannot be denied that the Information Technology Act is a 24-year-old law and since then several huge advancements concerning AI have occurred which require a dedicated law to be in place. The government of India is seeking to bring the Digital India Act towards the later part of the year which is likely to have a chapter dedicated to AI.[xvi]
VII. CONCLUSION AND WAY FORWARD
With the advent of AI, a distinct form of crimes are seen to evolve, mainly found on the darknet, which is: Cybercrime as a ‘service’ powered by artificial intelligence. This means, a person could hire an individual, mainly a hacker, to commit a cybercrime for him, and in the commission of that crime, the hacker would take assistance of AI.[xvii] In simple parlance, cybercrimes would be offered as a service, in commission of which AI would play a crucial role. As every coin has two sides, so does AI. It is similar to a knife which can either be used by a surgeon to save one’s life or by a thug to take one’s life. It cannot be denied that such situations raise alarm regarding the grave concerns raised by AI crimes.
Intellectual Property infringement seems to pose a great challenge for authors and inventors. Hence, as far as artificial intelligence vis-a-vis intellectual property rights are concerned, India needs to have a detailed and tailored legislation dedicated to AI in order to prevent IP infringement by these AI algorithms. There is a dire need to balance individual and societal interests while keeping in mind the requirement of creativity and innovation. The efforts being put in by global institutions as well as by the Indian government need to be fructified to be able to establish an efficient redressal mechanism.
[i] Amanda Helter, ChatGPT, TECHTARGET (March 29, 2024, 09:00 PM), https://www.techtarget.com/whatis/definition/ChatGPT.
[ii] Dr. Pavan Duggal, Artificial Intelligence & Law, Address at Friday Group, Mar. 7, 2024.
[iii] Ibid.
[iv] Ibid.
[v] Kyle Schnitzer & Priscilla DeGregory, ‘Humiliated’ NY lawyer who used ChatGPT for ‘bogus’ court doc profusely apologizes, NEW YORK POST, June 8, 2023, https://nypost.com/2023/06/08/humiliated-lawyer-apologizes-over-chatgpt-court-flub/amp/.
[vi] Dan Milmo, Two US lawyers fined for submitting fake court citations from ChatGPT, THE GUARDIAN, June 23, 2023, https://amp.theguardian.com/technology/2023/jun/23/two-us-lawyers-fined-submitting-fake-court-citations-chatgpt
[vii] Fredrikson & Byron PA, Judges Issue Guidance on Use of Generative Artificial Intelligence in Court Filings in the Wake of Attorneys Citing to Nonexistent Cases Created by ChatGPT, LEXOLOGY, Feb. 12, 2024, https://www.lexology.com/library/detail.aspx?g=cb234139-3a8a-4ff4-a4bd-ca8b3a9e8427.
[viii] Michael M. Grynbaum & Ryan Mac, The Times Sues OpenAI and Microsoft Over A.I. Use of Copyrighted Work, THE NEW YORK TIMES, Dec. 27, 2023, https://www.nytimes.com/2023/12/27/business/media/new-york-times-open-ai-microsoft-lawsuit.html
[ix] Dr. Pavan Duggal, Artificial Intelligence & Law, Address at Friday Group, Mar. 7, 2024.
[x] Greg Bensinger, ChatGPT launches boom in AI-written e-books on Amazon, REUTERS, Feb. 22, 2023, https://www.reuters.com/technology/chatgpt-launches-boom-ai-written-e-books-amazon-2023-02-21/
[xi] G.R.Raghavendra & Gurujit Singh, Can Artificial Intelligence (AI) Machine Be Granted inventorship in India?, 28 JIPR 123, 126 (2023).
[xii] UNODC, Ad Hoc Committee to Elaborate a Comprehensive International Convention on Countering the Use of Information and Communications Technologies for Criminal Purposes, Mar. 14, 2024, https://www.unodc.org/unodc/en/cybercrime/ad_hoc_committee/home.
[xiii] Dr. Pavan Duggal, Artificial Intelligence & Law, Address at Friday Group, Mar. 7, 2024.
[xiv] Illinois’ Artificial Intelligence Video Interview Act: What You Need to Know, Frost Brown Todd Attorneys, Jan. 13, 2020, https://frostbrowntodd.com/illinois-artificial-intelligence-video-interview-act-what-you-need-to-know/#:~:text=Illinois%20is%20the%20first%20state,for%20positions%20based%20in%20Illinois.
[xv] Niti Aayog, National Strategy for Artificial Intelligence, June 2018, https://www.niti.gov.in/sites/default/files/2023-03/National-Strategy-for-Artificial-Intelligence.pdf
[xvi] Digital India Act 2023: Revolutionizing Internet Regulation in India, Oct. 27, 2023, https://www.nextias.com/blog/digital-india-act/#:~:text=The%20Digital%20India%20Act%20(DIA)%20incorporates%20a%20range%20of%20crucial,the%20continuously%20evolving%20digital%20environment.
[xvii] Dr. Pavan Duggal, Artificial Intelligence & Law, Address at Friday Group, Mar. 7, 2024.