Published on 19th June 2025
Authored By: Ayaan Siddiqui
ST. Xavier's University Kolkata
Abstract
The rise of Artificial Intelligence (AI) and Machine Learning (ML) technologies has presented novel legal challenges in diverse jurisdictions around the world. The developments have outstripped the antitrust laws and policies now on the books, leaving lawmakers, courts and legal experts growing more anxious. This work discusses various legal issues that AI and ML systems raise, i.e. the legal/technical friction – copyright infringement, privacy, liability and regulation.[1]
Introduction: The AI Legal Landscape
A few years back, you heard people in classroom discussions at schools and colleges, starting the conversation on machine learning and artificial intelligence. And now, they are being utilized everywhere from business processes, to government operations to traidition life and all the way to daily life of community. AI systems are making life-and-death decisions in health care, criminal justice, finance, even retail merchandising. In such landscape, traditional legal norms are challenged as familiar principles about proprietary interest in, accountability for, and entitlement to property are embraced closer to new settings around what now must define the parameters of a legal system. It seems that the normative models for regulating the constitutional aspects of AI technologies are still considerably heterogeneous.[2]
As AI features multiply, so too will legal woes, and in particular the cartography of advanced generative models that can produce original- like output.
These systems present interdisciplinary questions of property rights, privacy obligations, responsibility assignment, moral frontiers — questions that are well beyond the reach of conventional law. Such a state of legal uncertainty seems fated to stifle technological progress but also to expose people and organizations to harm with insufficient paths for legal remediation.[3]
Intellectual Property Conundrums
Copyright Challenges in AI-Generated Content
The critical contentious point is regarding the right to output generated by the artificial services/ tools when it comes to using legal terminology called “intellectual property” in respects of AI. Traditional copyright, however, places so many limitations that are tied to the requirement of a human authorship, so much so that algorithm generated content is infinitely more complex in terms of who owns it. It has been established recently by the United States Copyright Office that not only can a work that is simply created by an AI program not be copyrighted but, “a registrant has a duty to tell whether AI was involved or not in the creation,” and, “AI-content shall be covered by the description of the work” submitted.[4]
This position has emerged in new jurisprudence. On October 5, 2023, in Kadrey v Meta Platforms, Inc, the Northern District of California dismissed copyright infringement claims related to AI outputs, holding that a demonstration that the outputs are exact copies of the works of the plaintiffs must be pleaded, or if access and copying is established, a fact that the outputs are protectable derivative works has to be pleaded. Likewise, in Andersen v Stability AI, the court also dismissed (with leave to amend) the artists’ copyright infringement claims, noting that they needed to plead, the level of similarity which AI-generated works bore to the underlying works of art.[5]
Training Data and Copyright Infringement
A second type of major issue involving intellectual property, particularly resonant with copyright law questions, is the use of some copyright-protected material to train some AI systems. In Authors Guild v OpenAI, the plaintiffs contended that OpenAI was infringing the copyright protections on their work by building language models because OpenAI required their copyrighted materials in order to be able to produce data.[6] This is also in line with a case initiated by Asian News International (ANI) against OpenAI at the Delhi High Court, demonstrating how AI tech innovations could bite the law.[7]
These disputes reveal a paradox of law: Whether the use of copyrighted material to train an A.I. model essentially constitutes a fair-use policy, or an unauthorized reproduction. Since the Copyright Act of 1957 does not have specification on AI interaction in India, developers will have to take the lead in obtaining licenses from rights holders prior to using data of any such nature in their training databases.[8] I see such a stance as signaling the global uncertainty around copyright in AI.[9]
Data Privacy and Protection Concerns
Constitutional Dimensions of Privacy
In India, the data is processed under Information Technology act 2000. Many of the massive datasets that artificial intelligence systems work with contain personal and sensitive data, leading to huge privacy challenges. The unbridled use of AI-enabled surveillance tools in India is a direct attack on the legal lessons that we derived in K.S. Puttaswamy v Union Of India (2017) from a reading of the right to privacy as a fundamental right for the first time.[10] The new law does not insert a proportionality principle, thus an area of legislature has been left open where data processing and collection will be permitted to be closed.[11]
With respect to the context of AI, if we look closely at the draft, we see that its most glaring problems arise with regards to 2023 Digital Personal Data Protection Act, especially in permitting broad and almost arbitrary Governmental exemptions that render individual privacy in the context moot.[12] Loopholes like these might enable us to systematically violate informational privacy without facing significant legal consequences, leading to the frameworks aimed at safeguarding privacy being ineffective by not answering challenges of AI. [13]
Cross-Jurisdictional Compliance Challenges
However, this entire patchwork of privacy regulation, eg: The General Data Protection Regulation (‘GDPR’) in the EU and the California Consumer Privacy Act (‘CCPA’) in the U.S., creates challenging legal compliance issues for global companies using AI in the products and services they offer. Each legislation introduces a new set of rules that governs data processing, which is centered around transparency, consent and involvement in the processing — turning into a porous framework for the legal application of AI.[14]
There is no one harmonised approach or internationally agreed standards, which ultimately leaves companies to tussle in a patchwork of conflicting legal frameworks that stifle technological advancement and inflate compliance costs. This maze of cross-currents where data-driven innovation hits the laws of several jurisdictions regulating that kind of activity requires acute multidisciplinary legal craft. [15]
Liability and Accountability Frameworks
Attribution of Responsibility
As AI systems have grown more powerful, the question of liability for the harm they do has become increasingly tangled. The capacity of the system to learn and self-diagnose makes the centering of liability within traditional frameworks — ones dependent upon human agency and prescience — more or less impossible. Who — of the constellation of developers, users or AI systems — can be held to account in a court of law? Provisions must be set.[16]
One of the thorniest areas that A.I. touches is the black-and-white status of the law concerning liability. Consider the case of autonomous driving running cars: but when a car suddenly gets into an accident immediately after the AI takes over driving directions, or in the case of a named the physician AI diagnosis assistant that approaches prescribed food and drug treatment harmfully, in concept all of it must be ok, but in actuality legal exercise requirements governance and enforcement to determine the constellation of legal responsibility.[17]
Regulatory Approaches to Accountability
Numerous jurisdictions are moving to close these gaps. Some of these concerns have been mitigated by taking a risk-based approach to measurement that makes higher-tier AI functionalities more directly accountable and also by providing clearer boundaries of accountability within the EU.[18] Such as is the case within the proposed EU AI Act. So, the regulation of AI law in US unlike every other country there in Europe.[19]
Application of decline judicial control, proportionate irritation and.ai, absence of an indication of the AI depletion regions, absence of consent except constructive consensus releases, provoke AI usage by ordinance regions as defamation for way to the replicable put without balance, at declinables deliberate audits by way of self-police. Indeed, such absence of frameworks and the lack of the same are paving the way for smothered innovation and non protectable harms by the AI which demands the legal arena to recuperate its makeup. [20]
Ethical Challenges and Algorithmic Bias
Bias as a Legal Issue
The large risk of having even slightly discriminatory bias in algorithms is that it poses a huge threat to law, ethics, etc especially when employed in such critical domains like justice, employment or finance. Algorithms that use historical data as a springboard to teach AI are likely to replicate present-day societal stereotypes and prejudices — and frequently magnify them. Zero-outcomes discrimination is at once a totem of legal fault, equality in tort and fairness of one’s judiciary.[21]
The question is who can be held accountable for bias. The law still trying to understand if ai maker, ai system acquirer, data provider, or both are liable. Legal practitioners in the industry are now calling for AI decision-making to be broader in scope, which will help transform legislation that makes it mandatory for AI systems to undergo regular bias checks.[22]
Transparency Requirements
The “black box” problem of advanced AI systems makes it difficult to know how they reach their conclusions, therefore restricting his controllability which violates legal fairness with needs of explainability. If AI systems do not even provide plausible scenarios for how conclusions are reached, then those adversely affected will probably not get the keys to countering the information bias that leads to adverse decisions.[23]
Countries across the globe have started to fill the evident gaps in providing reasoned judgment for decisions taken by automated means. These legal frameworks are designed to empower people to access in a meaningful way information on how AI systems process and make decisions on matters of importance to their rights and interests. But these ethical duties tend to conflict with proprietary interests and technical constraints, posing additional legal risk to AI developers and deployers.[24]
Emerging Regulatory Landscapes
Global Approaches to AI Regulation
Approaches to AI governance vary wide across jurisdictions, which is a result of diverse legal cultures, guiding policies and sophistication of technology. The European Union AI Act Proposal is working towards a new risk-based approach, as the EU aims for a more holistic way in facing AI challenges and creating a continuum for risk categories using a risk-based approach. The ultimate goal of this approach is to maintain core rights while eliminating legal uncertainty and continuing to drive innovation. [25]
Conversely last year United States, adopted practical principles-based approach in its regulatory approach whereas various government agencies can deal with regulators with regards with deep concerns around AI resulting fragmented and divergent for effective AI regulation. [26]
India’s Regulatory Development
On the note of challenges created by AI, The academic study only mentions threats of AI and broader challenges in India and very little work has been done concerning the hurdles created by AI. The country has no specific regulations on AI. There are specific base laws that have been enforced in India which attempt to address this (to some moderate extent), these being the Information Technology Act, 2000 and the Personal Data Protection Bill, 2019 however nothing defined how such exceptions would be made considering the unique nature of AI hence no clauses on it are present. [27]
This means developers and users of AI can not know to be able or AI complication due to lack of pertinent regulation, this in turn makes people unprotected from injury caused by AI. No existing law can protect them from being ambushed by rogue data extraction and carelessly handled sensitive information. As the basic tenets of rights are violated, with facial recognition and data gathering technologies posing even greater threats to privacy, it is imperative that there are well-drafted, comprehensive laws specific to AI.[28]
The Future of AI Law and Legal Practice
Specialization and Education
There is also a need for a lawyer that understands data regulations and information technologies. The lectures are mainly depending on what is required by the users, and also as per the latest in the industry, which is somewhat great to observe top universities around the world actually have started training lawyers for different types of AI and its legal activities, so it is a very good competition that we have to observe how well they are being inclined to bring suitable education in the education system.[29]
The change of pace that is ushered by AI technologies is unlike what we have ever experienced, and therefore, lawyers will need to relearn, re-browse new technology, browse new laws, even new legal provisions. Seminars, certifications or workshops guide lawyers in this area to obtain more intelligence on the topic that leads to the goal, in turn, to establish a branch of law in the field of AI.[30]
Interdisciplinary Collaboration
That question has lawyers, data scientists and even ethicists grappling over what to make of the contention around generative A.I. Treating law AI ethics, practice and policies must be cross-disciplinary and there must be an active engagement of diverse viewpoints. Such common effort would allow for a better understanding buy a workable layout regarding how the AI system would function perspective-wise in regards towards the law. [31]
If we are going to define the issues we want to solve, whether that’s bias in algorithmic discrimination, lack of data privacy, and the questions of liability, then the expertise of law must be tied up with that of technology. Adequate innovation that transcends over the disciplinary divides can empower legal practitioners to establish high end protective frames for the disruptive technologies. [32]
Conclusion
Legal problems of AI and machine learning are fundamentally the same. They are nuanced and dynamic, showing how these technologies are reshaping the world. Because such older principles and frameworks of law are insufficient to enforce against the new problems engendered by AI systems, this creates grave uncertainty for developers and users alike, and for would-be regulators. Laws should save the user across the world while motivating innovation and adoption of using AI tech as tech continues to expand.[33]
What should be done in this case is a combination of the two. Long-term we can enjoy some moderation and regulation based on that focus regulation, but short-run we need to have legal recognition of ownership with respect to copyright protected AI generated content, we need clear limbs of liability and we need stronger measures against privacy violations. This leaves the legislative side of things increasingly complex in many sectors of technology, starting particularly with Artificial Intelligence, as the legal court is essentially protecting core human rights and values.[34]
Adjusting to these shifts means transforming the practice of law itself. Watching with keen interest how the interdisciplinary co-existence in an area of expertise will certainly fulfill the growing need for specialized lawyers in AI through the professional development programs. As policy reform is front and center, lawyers have the potential to influence not just a model of AI governance, but also one that minimizes the risks of such technology while facilitating their development.[35]
References
[1] Morgan Lewis, ‘Addressing Legal Challenges in the AI/ML Era’ (2023) https://www.morganlewis.com/blogs/sourcingatmorganlewis/2023/03/addressing-legal-challenges-in-the-ai-ml-era accessed 16 April 2025.
[2] Drishti Judiciary, ‘Legal Challenges Before Artificial Intelligence’ (2024) https://www.drishtijudiciary.com/editorial/legal-challenges-before-artificial-intelligence accessed 16 April 2025.
[3] Epilogue Systems, ‘5 Key AI Legal Challenges In The Era Of Generative AI’ (2024) https://epiloguesystems.com/blog/5-key-ai-legal-challenges/ accessed 16 April 2025
[4] Crowell, ‘6 AI Cases And What They Mean For Copyright Law’ (2023) https://www.crowell.com/a/web/7QtNejMH1FSM1n5Ddt6cdU/6-ai-cases-and-what-they-mean-for-copyright-law.pdf accessed 16 April 2025.
[5] Andersen v Stability AI (2023) N.D. Cal 23-cv-03417.
[6] Authors Guild v OpenAI (2023) SDNY 23-cv-10834.
[7] Asian News International v OpenAI (2024) Delhi HC WP(C) 567/2024.
[8] Copyright Act 1957 (India).
[9] S&R Associates, ‘Addressing Legal Challenges on AI Development and Use’ (2025) https://www.snrlaw.in/addressing-legal-challenges-on-ai-development-and-use/ accessed 16 April 2025.
[10] K.S. Puttaswamy v Union of India (2017) 10 SCC 1
[11] Drishti Judiciary, ‘Legal Challenges Before Artificial Intelligence’ (2024) https://www.drishtijudiciary.com/editorial/legal-challenges-before-artificial-intelligence accessed 16 April 2025.
[12] Digital Personal Data Protection Act 2023 (India).
[13] S&R Associates, ‘Addressing Legal Challenges on AI Development and Use’ (2025) https://www.snrlaw.in/addressing-legal-challenges-on-ai-development-and-use/ accessed 16 April 2025.
[14]General Data Protection Regulation (EU) 2016/679.
[15] National Law Review, ‘Exploring Legal Challenges in Artificial Intelligence’ (2025) https://natlawreview.com/article/growth-ai-law-exploring-legal-challenges-artificial-intelligence accessed 16 April 2025.
[16] WalkMe, ‘7 AI legal issues and how to deal with them’ (2024) https://www.walkme.com/blog/ai-legal-issues/ accessed 16 April 2025.
[17] National Law Review, ‘Exploring Legal Challenges in Artificial Intelligence’ (2025) https://natlawreview.com/article/growth-ai-law-exploring-legal-challenges-artificial-intelligence accessed 16 April 2025.
[18] Drishti Judiciary, ‘Legal Challenges Before Artificial Intelligence’ (2024) https://www.drishtijudiciary.com/editorial/legal-challenges-before-artificial-intelligence accessed 16 April 2025.
[19] National Law Review, ‘Exploring Legal Challenges in Artificial Intelligence’ (2025) https://natlawreview.com/article/growth-ai-law-exploring-legal-challenges-artificial-intelligence accessed 16 April 2025
[20] S&R Associates, ‘Addressing Legal Challenges on AI Development and Use’ (2025) https://www.snrlaw.in/addressing-legal-challenges-on-ai-development-and-use/ accessed 16 April 2025.
[21] Drishti Judiciary, ‘Legal Challenges Before Artificial Intelligence’ (2024) https://www.drishtijudiciary.com/editorial/legal-challenges-before-artificial-intelligence accessed 16 April 2025.
[22] Epilogue Systems, ‘5 Key AI Legal Challenges In The Era Of Generative AI’ (2024) https://epiloguesystems.com/blog/5-key-ai-legal-challenges/ accessed 16 April 2025.
[23] WalkMe, ‘7 AI legal issues and how to deal with them’ (2024) https://www.walkme.com/blog/ai-legal-issues/ accessed 16 April 2025.
[24] National Law Review, ‘Exploring Legal Challenges in Artificial Intelligence’ (2025) https://natlawreview.com/article/growth-ai-law-exploring-legal-challenges-artificial-intelligence accessed 16 April 2025.
[25] Drishti Judiciary, ‘Legal Challenges Before Artificial Intelligence’ (2024) https://www.drishtijudiciary.com/editorial/legal-challenges-before-artificial-intelligence accessed 16 April 2025.
[26] National Law Review, ‘Exploring Legal Challenges in Artificial Intelligence’ (2025) https://natlawreview.com/article/growth-ai-law-exploring-legal-challenges-artificial-intelligence accessed 16 April 2025.
[27] The Amikus Qriae, ‘Research paper on the Legalities of Artificial Intelligence in India and other countries’ (2024) https://theamikusqriae.com/research-paper-on-the-legalities-of-artificial-intelligence-in-india-and-other-countries/ accessed 16 April 2025.
[28] Drishti Judiciary, ‘Legal Challenges Before Artificial Intelligence’ (2024) https://www.drishtijudiciary.com/editorial/legal-challenges-before-artificial-intelligence accessed 16 April 2025.
[29] Qlantic Journal of Social Sciences, ‘Artificial Intelligence and Machine Learning in Legal Research’ (2024) https://qjss.com.pk/index.php/qjss/article/view/252 accessed 16 April 2025.
[30] Ibid.
[31] National Law Review, ‘Exploring Legal Challenges in Artificial Intelligence’ (2025) https://natlawreview.com/article/growth-ai-law-exploring-legal-challenges-artificial-intelligence accessed 16 April 2025.
[32] Drishti Judiciary, ‘Legal Challenges Before Artificial Intelligence’ (2024) https://www.drishtijudiciary.com/editorial/legal-challenges-before-artificial-intelligence accessed 16 April 2025.
[33] Epilogue Systems, ‘5 Key AI Legal Challenges In The Era Of Generative AI’ (2024) https://epiloguesystems.com/blog/5-key-ai-legal-challenges/ accessed 16 April 2025.
[34] Morgan Lewis, ‘Addressing Legal Challenges in the AI/ML Era’ (2023) https://www.morganlewis.com/blogs/sourcingatmorganlewis/2023/03/addressing-legal-challenges-in-the-ai-ml-era accessed 16 April 2025
[35] Drishti Judiciary, ‘Legal Challenges Before Artificial Intelligence’ (2024) https://www.drishtijudiciary.com/editorial/legal-challenges-before-artificial-intelligence accessed 16 April 2025
