I. ABSTRACT
The Article emphasizes the symbiotic relationship between artificial intelligence (AI) and human creativity, citing Fei-Fei Li's view that AI amplifies human ingenuity. It traces the evolution of generative AI, particularly neural networks, highlighting their role in creating new content based on learned patterns. The emergence of generative AI raises legal questions about copyright ownership, especially concerning works created without human intervention. The discussion extends to the risks of deepfake AI, its potential for misinformation, and the lack of legal safeguards. Additionally, the text explores the ethical concerns of data privacy and the societal impacts of AI on various industries, advocating for transparent and effective legal solutions.
II. INTRODUCTION
“Artificial intelligence is not a substitute for human intelligence; it is a tool to amplify human creativity and ingenuity”, [ii]famous words by Fei-Fei Li, an American computer scientist at Stanford is a pioneering expert on generative AI and machine learning. Generative AI finds it roots in the early development of machine learning[iii], when scientists and researchers in the late 1950s started to explore the potential benefit of machine learning and how humans can automate their routine tasks, machines can learn from a fixed pattern of results and much more. It was during the 1990s and 2000s, that the field of machine learning, and artificial intelligence went on a fast-track because of the advanced hardware and digital data become more widely available.
Generative AI is regarded as one of the unique kinds of Artificial intelligence that focusses on creating new content, which include all text, images, audio, video, based on specific patterns that it has learned from the data and content fed to it during its training. To generate accurate and relevant information, generative AI model is based and must be trained on enormous volumes of data using deep learning, neural networking to produce answer to questions, write stories, produce mages, audio, source code of any description, all based on brief text inputs or “prompts” given by the user[iv].
Earlier, the ownership of copyright in computer generated works was not in question because the hardware and software were merely just a tool that supported the creative process, just like a computer aided software or very similar to pen and paper. But with the creation and advent of the latest types of artificial intelligence, the computer program has no longer remained a tool to support the creative process but has emerged as an independent tool of producing creating works in various formats and types without ever needing the aid of human intervention. While most domestic laws, recognize creative and original works produced or guided by a human creator behind it, will qualify for copyright protection if they are original. [v]This clause will raise serious lacunae in our contemporary laws and it becomes very important for the Indian law to fill gaps and answer important questions related to who will finally own the work generated by AI, the user or the Owner of the AI? How original content will be protected from the use of AI at Educational institutions and Academic spheres? How to stop the spread of misinformation using AI tools such as deepfakes and Gen-AI? What shall be the different criteria’s to be able to distinguish between original work and AI generated work?
III. UNDERSTANDING AI-GENERATED DEEP-FAKES AND THEIR CONSEQUENCES
Deepfake AI is on the types of artificial intelligence used to create very human-alike, realistic, and convincing images, audio and video hoaxes. The term itself describes both the technology as well as the resultant pseudo content and is a culmination of deep learning and unreal product. The biggest danger evident from the use of deepfakes is their ability to spread false information that may appear to come from trusted or verifiable sources, but they aren’t even close to it. [vi]For example, in 2022 a deepfake video was released of Ukraine president Volodymyr Zelenskyy asking his troops to surrender to Russian army. And in India recently, a deepfake video of Indian celebrity Rashmika Mandanna had surfaced on the internet in November 2023 earlier on various social media platforms, which raised serious questions about the fair use of such AI tools and their resultant products and following the incident, Delhi police filed a case under sections 465 and 469 relating to forgery and harming the reputation of a women under IPC,1860 along with privacy violation and identity theft under Sections 66 of IT Act, 2000.
Deep-Fakes uses two different sets of algorithms – a generator and a discriminator – to create and refine fake content, while the former is used to train the AI on the data set, the latter analyzes the distinction between the realistic or fake the initial version of the content. Deep-fakes videos are created either using the original source of the data or they can swap the person’s face onto a video of another individual. [vii]
There is almost little to no protection given for using deep-fakes, until they violate a pre-existing law regarding identity theft under Section 66C, privacy violation under Section 66E of the Information Technology Act, 2000 or under Sections 465 (forgery), 469 (forgery for purpose of harming reputation), 499 (defamation), etc just like in the case of Rashmika Madanna, which is just one of many cases at hand that explains this gap in law. This regulatory gap leaves victims vulnerable in an increasingly digitized world. The lack of specific legislation is partly attributable to a widespread lack of awareness among the public regarding the capabilities and dangers of deepfake technology.
Data is considered to be a very asset for artificial intelligence to train on, as they use large amounts of data sets to train themselves, and the problems lie therein, as the data used by them is easily available on public domain and websites such as social media sites, blogs, publicly available articles and so on. This data is usually used without the permission of the user who created that data in the first place which poses a real problem to privacy as well as the security of the data of that person.
These AI companies are involved in the act of gross violation of our privacy as well digital security by scrapping of huge sets of data [viii]which includes texts, images, videos, speech and voice, as well personal data available publicly off the Internet for training their artificial intelligence model, and as held under the landmark judgement of KS Puttaswamy v. Union of India, the apex court specially recognized the right to privacy under the larger ambit of Article 21 of the Indian Constitution.
Generating new and creative content based on large data sets makes it a boon in many industries and fields of work. The efficiency and speed of such technology can help create us content in less time and deliver promising results. This may have a direct or indirect impact on the requirement of human labor and need for man-power, which might raise some eye-brows about suspected job displacement followed by income inequality. Gen-AI in the field of healthcare is used to analyze medical images and assist doctors in making diagnoses, in the field of finance it has shown potential to detect fraud and identify investment opportunities, and seems to be a good fit with it as it deals with vast amounts of data, which is what finance relies on, in entertainment industry, it helps create personalized recommendations and targeted content based on individual preferences, and much more.
IV. LEGAL OPTIONS AND SOLUTIONS
Works that have very little to no human contact can be handled in either one of two ways under copyright law. For works produced by computers (GenAI), it can either refuse copyright protection or assign authorship to the program's developer and its owner.
On one hand, providing copyright especially in works generated by artificial intelligence has never been specifically prohibited, but there might be indications that the laws of many such countries are not comprehending to non-human copyright. For example, in United States, for example the Copyright Office has declared that it will “register an original work of authorship, provided that the work was created by a human being.” “This stance flows from case law (e.g. Feist Publications v Rural Telephone Service Company, Inc. 499 U.S. 340 (1991)) which specifies that copyright law only protects “the fruits of intellectual labor” that “are founded in the creative powers of the mind. “Similarly, in a recent Australian case (Acohs Pty Ltd v Ucorp Pty Ltd), a court declared that a work generated with the intervention of a computer could not be protected by copyright because it was not produced by a human.[ix]
While the latter exist, which is giving the due authorship to the programmer, is evident in a few countries such as the Hong Kong, India, and the UK. This one is explicitly mentioned under the UK copyright law, under Section 9(3) of the Copyright, Designs and Patents Act (CDPA), which states as under:
“In the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken.”
Such options often lead to ambiguity of who the law would genuinely recognize to be the person making the arrangements for the work to be generated. Should the law recognize the contributions of the programmer or the proprietor of the program? Should there be no differentiation between the programmer or the user? When determining who wrote a computer game, the Court of Appeal in the English case of Nova Productions v. Mazooma Games [2007] EWCA 219 ruled that a player's contribution “is not artistic in nature and he has contributed no skill or labour of an artistic kind.” Thus, taking user behaviour into account on a case-by-case basis could be one way to solve the issue.
Thus, it becomes very important that we shall look back into the loopholes and fallacies in the law to address these pressing concerns as AI and its implications are here to stay and will not be-gone as it is penetrating deeper into our everyday lives and our daily routine. There should be more transparent and effective form of mode to address such concerns of such disputes and shall be decided on a case-to-case basis. We should also be able to determine accountability and responsibility in cases of copyright infringement, violation of privacy, and cases of originality threat.
[i] The author is a student at the Rajiv Gandhi National University of Law, Patiala.
[ii] Alex Galt, Unlocking A world of possibilities: How AI is revolutionizing real estate and empowering your life choices, Forbes https://www.forbes.com/sites/forbesbusinesscouncil/2023/12/18/unlocking-a-world-of-possibilities-how-ai-is-revolutionizing-real-estate-and-empowering-your-life-choices/?sh=7d07676961bc.
[iii] A simple guide to the history of Generative AI Bernard Marr, https://bernardmarr.com/a-simple-guide-to-the-history-of-generative-ai/.
[iv] What is Generative AI? Artificial intelligence that creates InfoWorld, https://www.infoworld.com/article/3689973/what-is-generative-ai-artificial-intelligence-that-creates.html.
[v] Andres Guadamez, Artificial Intelligence and copyright WIPO, https://www.wipo.int/wipo_magazine/en/2017/05/article_0003.html
[vi] What is Deepfake AI? A definition from TechTarget WhatIs, https://www.techtarget.com/whatis/definition/deepfake
[vii] Id
[viii] Ben Canner, The dangers of data scraping: Do you know what’s out there? Best Endpoint Protection Security (EPP) Tools, Software, Solutions & Vendors (2021), https://solutionsreview.com/endpoint-security/the-dangers-of-data-scraping-do-you-know-whats-out-there/
[ix] Supra note 4
Comments