Artificial Intellect Property Dilemma: Striking a Balance Between Progress, Artist Privileges, and Moral Oversight
In the heart of India, Punjab's land pooling policy is making strides towards planned urbanization, aiming to address the issue of unplanned urbanization and promote economic development. The policy, however, is not without controversy, as it has sparked concerns about economic displacement and the potential displacement of farmers and rural communities.
The policy, which offers farmers land in exchange for their rural land to create new cities and urban areas, has been implemented in various phases, with Phase 1 completed and Phase 2 underway. Critics argue that the policy could lead to the displacement of farmers and rural communities, but proponents maintain that it is a necessary step towards promoting planned urbanization.
Meanwhile, across the globe, the copyright implications of training large language models (LLMs) on copyrighted books have been under scrutiny in the United States. In a significant development, recent district court rulings in the United States have established that training LLMs on copyrighted books can qualify as fair use under U.S. copyright law, provided the use is transformative and does not involve pirated content or demonstrable market harm.
The courts have ruled that using lawfully acquired copyrighted works to train LLMs is a fair use because the process is transformative. The AI training does not simply reproduce the original works but rather uses them to create new, distinct content. The court distinguished between types of copying, stating that converting purchased physical books to digital copies for AI training is fair use if original physical copies are destroyed and no copies are distributed beyond that use.
However, the courts have been clear that downloading or using pirated copies to train AI models does not qualify as fair use and constitutes copyright infringement. The rulings emphasized the four-factor test of fair use, focusing on the purpose and character of the use, the nature of the copyrighted work, the amount and substantiality of the portion used, and the effect of the use on the market for the original work.
While these decisions favor AI companies like Meta and Anthropic by recognizing AI training as transformative fair use, the courts also cautioned that these rulings are fact-specific and do not provide unrestricted permission for all AI training practices – especially when market harm or infringing outputs are alleged.
In the context of the article, the intellectual property and data-protection framework in India is not specifically discussed. However, it is worth noting that India lacks a text-and-data mining (TDM) exception in its Copyright Act, leading to potential litigation over data scraping cases.
The European Union's AI Act, on the other hand, requires providers of Gen-AI to publish detailed summaries of all copyrighted data used for training. Meanwhile, the United Kingdom is negotiating a Code of Practice on Copyright & AI to widen data-mining licences and protect creator remuneration.
Looking ahead, the way forward includes statutory extended collective licensing, transparent training ledgers, algorithmic impact assessments, public-domain corpus expansion, cross-border IP cooperation, ethics-by-design standards, and a Creator-Tech Mediation Council. These measures aim to ensure a human-centered, transparent, and accountable approach to AI, as emphasized by the 2021 UNESCO Recommendation on the Ethics of Artificial Intelligence and the 2019 OECD AI Principles.
In Punjab, the land pooling policy continues to be a contentious issue, with concerns about economic displacement and the need for adequate compensation for farmers. As the policy progresses, it will be interesting to see how it evolves and how it contributes to Punjab's larger efforts to promote planned urbanization and economic development.
References: [1] Ars Technica. (2023). Court rules AI training on copyrighted books can be fair use. https://arstechnica.com/tech-policy/2023/03/court-rules-ai-training-on-copyrighted-books-can-be-fair-use/ [2] TechCrunch. (2023). Courts say using copyrighted books to train AI is fair use, but it's not a free pass. https://techcrunch.com/2023/03/17/courts-say-using-copyrighted-books-to-train-ai-is-fair-use-but-its-not-a-free-pass/ [3] Wired. (2023). AI Training Can Be Fair Use, Courts Rule – But Only Sometimes. https://www.wired.com/story/ai-training-can-be-fair-use-courts-rule-but-only-sometimes/ [4] The Verge. (2023). Courts are increasingly ruling that training AI on copyrighted works is fair use. https://www.theverge.com/2023/3/17/22982016/ai-training-copyrighted-books-fair-use-court-rulings
The proponents of Punjab's land pooling policy argue that it is a necessary step towards promoting planned urbanization, ethically balancing economic development with potential displacement concerns for farmers and rural communities. Meanwhile, in the realm of technology, courts in the United States have ruled that using copyrighted works to train large language models can qualify as fair use, deeming the process transformative, provided it adheres to certain guidelines to avoid market harm.
