OpenAI and Google Request Copyright Exceptions for AI Training: A Bold Move for Innovation or Creative Industry's Nightmare?

The intersection of artificial intelligence (AI) and copyright law represents a rapidly evolving and complex area of debate. Recently, tech giants like OpenAI and Google have petitioned the U.S. Copyright Office for exceptions that would allow their AI models to be trained on copyrighted content without obtaining licenses, arguing that such practices fall under "fair use". This request has sparked significant discussion regarding the balance between technological advancement and the rights of content creators. The concept of fair use in the U.S. allows for limited utilization of copyrighted material without explicit permission, often involving a nuanced balance of several factors such as the purpose of use and its effect on the market for the original work. These considerations are at the forefront of the current debate on whether AI technologies should have broader access to copyrighted content for training purposes. OpenAI, in particular, emphasizes that this issue extends beyond business and into national security, reflecting the strategic importance they place on maintaining a leading edge in AI capabilities. For more details on this request, you can visit this [link](https://www.niemanlab.org/reading/openai‑and‑google‑ask‑the‑government‑to‑let‑them‑train‑ai‑on‑content‑they‑dont‑own/).

The request by OpenAI and Google reflects broader challenges the AI industry faces as it grapples with the need for vast and comprehensive datasets. By accessing copyrighted material, these companies argue they can develop more sophisticated AI models that drive innovation and societal benefits. Conversely, there are strong opposition voices, particularly within the creative industry, concerned about market devaluation and loss of incentives for original creations if such copyright exceptions are granted. This tension highlights the need to adapt current legal frameworks to the realities of AI development while safeguarding the economic and cultural value of intellectual property. Such adaptations are complicated by ongoing copyright lawsuits against companies like OpenAI, which test the limits and applicability of fair use in AI training. For a deeper dive into these legal battles and their implications, explore this [reading](https://www.niemanlab.org/reading/openai‑and‑google‑ask‑the‑government‑to‑let‑them‑train‑ai‑on‑content‑they‑dont‑own/).

In recent developments around artificial intelligence, two tech giants, OpenAI and Google, have jointly advanced a controversial request to the U.S. Copyright Office, seeking an exception to copyright law. This move aims to circumvent the need for obtaining licenses to use copyrighted material in AI model training. They argue this falls under 'fair use'—a legal doctrine permitting limited use of copyrighted material without the need for permission, provided the use meets certain conditions such as non‑commercial intent or educational use. OpenAI asserts that unrestricted access to rich textual datasets is not just a commercial advantage but a national security imperative, claiming that enhanced AI capabilities are crucial for maintaining the U.S.'s global leadership in technology [source](https://www.niemanlab.org/reading/openai‑and‑google‑ask‑the‑government‑to‑let‑them‑train‑ai‑on‑content‑they‑dont‑own/).

The request by OpenAI and Google highlights significant tensions in how copyright law intersects with digital and technological innovation. Training large language models requires massive and diverse datasets, often comprising texts that are under copyright protection. The tech companies argue that acquiring individual licenses for these volumes of content is impractical and prohibitively expensive. By invoking 'fair use,' they propose that the training process does not detract value from the original works, as the output AI models transform the data inputs into new, highly functional systems rather than replicating the content verbatim.

Support for this proposal varies, reflecting a broader divide between innovation advocates and copyright traditionalists. Proponents suggest that to remain competitive, especially against emerging global AI leaders from places like China, U.S. companies need broader leeway in handling copyrighted works. Some experts argue that the current copyright framework imposes barriers that could inhibit technological progress and competitiveness [source](https://www.niemanlab.org/reading/openai‑and‑google‑ask‑the‑government‑to‑let‑them‑train‑ai‑on‑content‑they‑dont‑own/). Nevertheless, this stance is met with significant resistance from several stakeholder groups, including authors, publishers, and Hollywood creatives, who fear that such an exemption might undermine the economic vitality of creative industries and devalue their intellectual property.

The debate around this initiative is also punctuated by a series of legal battles, including lawsuits filed by news outlets and authors against OpenAI for alleged copyright infringement. These cases challenge the elasticity of 'fair use' in the context of AI and test whether transformative purposes of technology applications justify the use of protected content without explicit consent. Critics argue that granting a blanket exception could set a precedent that erodes fundamental copyright protections, leading to a broader chilling effect on creative content generation [source](https://www.niemanlab.org/reading/openai‑and‑google‑ask‑the‑government‑to‑let‑them‑train‑ai‑on‑content‑they‑dont‑own/).

The pursuit of this copyright exception by OpenAI and Google poses critical questions about the future balance between innovation and intellectual property rights. As the U.S. Copyright Office deliberates on these requests, the outcome could either accelerate AI innovation by making vast data resources readily available for training, or reinforce protections for creative industries, ensuring that any AI development process respects existing copyright laws and contributes to a fair economic ecosystem for content creators. This decision will undoubtedly influence the competitive landscape of AI technology, both domestically and internationally [source](https://www.niemanlab.org/reading/openai‑and‑google‑ask‑the‑government‑to‑let‑them‑train‑ai‑on‑content‑they‑dont‑own/).

The concept of 'fair use' holds significant meaning within the realm of copyright law, offering a pathway for limited usage of copyrighted material without obtaining explicit permission from the owners. This doctrine plays a critical role in the ongoing discussions about AI development, particularly in the context of companies like OpenAI and Google, who have argued for its application in training AI models. In a notable move, these tech giants have requested an allowance to utilize copyrighted content without needing licenses, citing "fair use" as the legal basis for their plea. The argument is rooted in the requirement for access to vast and varied textual data to train AI models effectively, a process that is pivotal to advancing AI technology and maintaining a competitive edge globally. However, this request has sparked a contentious debate, as it challenges traditional notions of copyright protection and raises questions about the potential impact on content creators and copyright holders. For further insights, readers can explore the detailed article on NiemanLab discussing OpenAI and Google's proposal and its broader implications.

In the fast‑evolving landscape of artificial intelligence, a pressing debate has emerged concerning the copyright exception for AI training. Proponents like OpenAI and Google argue for a reinterpretation of copyright standards, where using copyrighted content can fall under 'fair use'. They emphasize that this change is vital for maintaining the United States' competitive edge in the global AI arena. Both companies have reached out to the U.S. Copyright Office, underscoring that the access to diverse and extensive datasets is not merely a technological need but a question of national security .

The argument for copyright exception pivots on the transformative nature of AI training. Companies like OpenAI assert that their use of copyrighted material does not merely replicate existing works but rather transforms it into new, innovative outputs, such as sophisticated AI models. This transformation is seen as beneficial to society by accelerating technological advancements and offering competitive advantages over international rivals. Additionally, with global players like China's DeepSeek gaining ground, the urgency for the U.S. to empower its AI sector becomes apparent, as noted in several technology reports .

Proponents also highlight that existing copyright laws are ill‑suited for the nuances of modern AI practices. The current legal framework can complicate access, elongate negotiation periods with copyright holders, and place U.S. companies at a strategic disadvantage. Experts at Andreessen Horowitz, a prominent venture capital firm, have pointed out the necessity for clearer guidelines to navigate AI model training under the fair use doctrine. They argue that a refined legal approach could facilitate the deployment of AI innovations while ensuring ethical practices .

Economically, a copyright exception could stave off potential stagnation in AI development by enabling seamless training model iterations. However, critics warn that such moves could economically impair the creative sectors by diluting the value of copyrighted works. The dichotomy between technological progress and creative protection lies at the heart of this debate, with each sector asserting the centrality of their role to national interests. Amidst these tensions, the Copyright Office is tasked with balancing the need for innovation against the rights and incentives for traditional content creation .

Those objecting to the copyright exception argue that allowing AI firms to bypass traditional licensing agreements with copyright holders would lead to an erosion of the economic value of creative works. The core concern is that by permitting technology companies to use proprietary content without payment or proper compensation, the market value of such content could be significantly diminished. This could disincentivize content creation, as authors and artists could see their works exploited without sharing in the commercial benefits. As such, this act could undermine the very foundation of creative industries that rely heavily on the protection and monetization of intellectual property [source](https://www.hollywoodreporter.com/business/business‑news/hollywood‑pushes‑back‑openai‑google‑argument‑copyright‑1236166626/).

Furthermore, the fear of exploitation extends beyond economic critique. Many in the creative industry voice apprehension that without proper licensing, AI models could replicate content in ways that misrepresent the original creators' intentions or strip the content of its contextual value. Such a scenario could lead to ethical quandaries where AI outputs are considered transformative yet potentially disrespect the integrity of the source material. Critics argue for the necessity of maintaining attribution and ensuring that AI development does not come at the expense of original creators' rights and reputations [source](https://medium.com/vertical‑bar‑media/google‑openai‑assert‑fair‑use‑in‑ai‑training‑legal‑battle‑vbm‑90c5f41fc914).

Some experts also argue that providing this exception could set a dangerous precedent within the legal framework, effectively devaluing copyrighted works on a massive scale. This legal loophole could embolden other tech firms to seek similar exemptions, destabilizing the balance of copyright law and presenting a slippery slope towards the widespread appropriation of artistic content without just compensation. This concern is echoed by Hollywood's creative community, who have been vocal in their opposition, highlighting the long‑term risks to cultural industries [source](https://www.indiewire.com/news/business/ai‑open‑letter‑openai‑google‑copyright‑exemptions‑1235104394/).

There is also a significant pushback from economic and cultural sectors against the notion that national security and technological advancement should override property rights and market ethics. Critics argue that granting such exceptions could prioritize large technology firms over smaller creative enterprises, exacerbating issues of inequity in the tech landscape. The principle of "fair use" is seen by opponents as being stretched beyond its traditional bounds to serve the interests of a powerful few, thus requiring more robust and considerate legal interpretation [source](https://www.losangelesblade.com/2023/11/09/hollywood‑unites‑against‑openai‑google‑copyright‑exemption‑request‑an‑industrial‑proclamation/).

The reactions to OpenAI and Google's request for copyright exceptions to train AI on proprietary content have been diverse, reflecting the high stakes involved. Many in the creative industry, especially Hollywood, have voiced strong opposition, fearing that such exceptions could undermine the value of their work. References to an open letter by over 400 professionals emphasize their concern that these exemptions could create unfair advantages for tech giants while diminishing the worth of creative labor and its economic contribution .

Meanwhile, supporters argue that accessing copyrighted materials is essential for AI innovation and maintaining the United States' competitive edge in the global market. Proponents suggest that AI development could stall without such resources, potentially affecting national security and technological leadership . They argue that a balance can be struck with careful regulation and transparent use of data, mitigating possible adverse effects on content creators.

The legal landscape also plays a crucial role in shaping public opinions. Lawsuits against OpenAI and Google highlight the ongoing tensions and uncertainties surrounding AI's use of copyrighted content. Despite these challenges, some argue that clarifying legal frameworks to support AI development without eroding intellectual property rights is crucial for future technological advancements .

Furthermore, international competitiveness adds another layer to the debate. As global AI contenders such as China's DeepSeek push forward, the pressure mounts for U.S. companies to secure large datasets to stay relevant in the tech race. The discussion is not just about national innovation; it touches on broader geopolitical dynamics and the role of government in fostering a favorable environment for AI advancement .

The intersection of national security and AI development has become increasingly pronounced as global competitiveness intensifies. Leading tech companies like OpenAI and Google are advocating for adjustments in copyright law, arguing that the ability to use a wide array of data for AI training is not only a technological imperative but a national security concern. By gaining access to a diverse dataset, the U.S. could potentially maintain its leadership in AI innovation, which is crucial in the context of geopolitical competition. Countries like China are intensifying their advancements in AI, putting pressure on the U.S. to ensure its technological capabilities are not surpassed. OpenAI's stance [1](https://www.niemanlab.org/reading/openai‑and‑google‑ask‑the‑government‑to‑let‑them‑train‑ai‑on‑content‑they‑dont‑own/) highlights this strategic necessity.

Moreover, the argument for using copyrighted material without traditional constraints pivots on the idea of "fair use" as a means to bolster national defense capabilities. Fair use, a cornerstone of U.S. copyright law, allows limited use of copyrighted materials without express permission, provided certain conditions are met. In this context, OpenAI posits that having the most advanced AI systems is vital for national security, suggesting a direct link between technological superiority and the ability to safeguard national interests. The notion that AI development is intertwined with national security interests suggests that any regulatory decisions have far‑reaching implications, echoing through sectors of defense, intelligence, and economic competitiveness [1](https://www.niemanlab.org/reading/openai‑and‑google‑ask‑the‑government‑to‑let‑them‑train‑ai‑on‑content‑they‑dont‑own/).

In the rapidly evolving landscape of technology, legal and regulatory challenges present significant hurdles for artificial intelligence (AI) developers. The ongoing debate around copyright law reveals a pressing issue with no easy resolution. OpenAI and Google have requested a special exception from the U.S. Copyright Office to train their AI models using copyrighted materials without acquiring licenses, citing the principle of "fair use." This request underscores the tension between technological advancement and intellectual property rights, highlighting a profound question: How can we balance the need for innovation with the rights of creators? Learn more about this request.

The legal framework surrounding the "fair use" doctrine plays a critical role in this challenge. Fair use allows limited use of copyrighted content without explicit permission, considering factors like purpose and effect on the market. However, the application of this doctrine to AI training remains contentious. Proponents argue that using copyrighted materials transforms them into a new creation—a machine learning model—thus qualifying for fair use. Others, especially copyright holders, fear this could undermine the commercial value of their work, potentially leading to decreased investment in creative industries.

The arguments for granting an exception largely revolve around national interests and the global AI race. OpenAI's stance on national security is a strategic move to emphasize the potential benefits of expedited AI innovation. They assert that the competitive edge ensured by access to vast datasets is crucial not only for technological leadership but also for national security matters. This perspective is bolstered by experts who argue that the slowing down of innovation due to restrictive copyright laws could pose risks not just economically but also politically on a global scale.

Conversely, the opposition, including Hollywood creatives and other copyright holders, starkly opposes such exceptions. Their concerns include the economic impact on creative sectors, intellectual property rights erosion, and the potential setting of a precedent where creative works are devalued Explore more on the creative industry's stance. Lawsuits filed against AI companies emphasize fears about misuse of copyrighted content. These lawsuits aim to address whether current copyright doctrines can accommodate the nuances of AI technology.

The future of these legal challenges is uncertain, yet they will undoubtedly shape the landscape of AI development and copyright law. If OpenAI and Google succeed, it could pave the way for new legislative frameworks that better align with the fast‑paced advancements in AI. However, such a shift could also exacerbate tensions between technology companies and content creators, sparking further debate on how best to align innovation with ethical considerations. The outcomes from these discussions and legal battles will likely influence not just policy in the U.S. but also set precedents for international regulations, affecting global competitiveness and the ethics of AI development.

The request from OpenAI and Google to relax copyright laws for AI training models is poised to result in profound economic implications, affecting various sectors differently. If the U.S. Copyright Office grants the exception, AI companies could gain unfettered access to enormous troves of data, which would likely accelerate the development of sophisticated AI technologies. Such advancements could positively impact the tech industry by enhancing capability, efficiency, and global competitiveness. Indeed, U.S.-based AI developments might see a surge, reinforcing national security and keeping pace with international counterparts, such as China, which are quickly advancing in AI technology .

However, this potential economic growth for the tech industry could come at a stark cost to the creative sectors. Content creators and copyright holders fear significant economic setbacks resulting from the unauthorized use of their intellectual property. If AI companies can freely utilize copyrighted works without compensation, this might reduce the economic value of these works, leading to decreased investment in new content creation . This scenario raises concerns that the creative industry's economic and cultural contributions could be considerably undermined.

Conversely, refusing the exception could suppress technological innovation within the United States, stalling advancements as AI companies wrestle with the increasing cost and complexity of obtaining licenses. In a worst‑case scenario, this could damage the country's standing in the global AI race, weakening national security and potentially leading the tech industry to fall behind international competitors willing to adopt more permissive data practices .

Ultimately, decision‑makers will need to consider a complex blend of economic factors, balancing the need for technological innovation with protection for the creative economy. A nuanced approach might include facilitating negotiations for licenses or incentivizing the development of alternative datasets for AI training. Aligning the expansion of AI capabilities with the economic health of the creative industries will be crucial in maintaining a healthy, innovative, and competitive market structure .

The pursuit to train AI models on copyrighted materials without obtaining permission has sparked significant social and ethical implications. On a societal level, this move by companies like OpenAI and Google may alter perceptions about the rights of content creators versus technological advancement. If granted, the request could potentially devalue individual creativity, as it challenges the traditional idea of respecting intellectual property. Content producers might feel that their contributions are being exploited without fair compensation or recognition, potentially leading to a chilling effect on artistic and literary endeavors. Furthermore, such actions could exacerbate existing tensions between the tech giants and content‑producing industries, fostering a narrative where innovation prioritizes corporate gain over individual rights and ethical considerations. Public sentiment is likely to become increasingly divided, as some advocate for rapid technological advancement while others champion the protection of creative rights [source](https://www.niemanlab.org/reading/openai‑and‑google‑ask‑the‑government‑to‑let‑them‑train‑ai‑on‑content‑they‑dont‑own/).

Ethically, the ramifications of using copyrighted materials without proper licensing reach into the core values of fairness and respect for intellectual property. Granting such an exception could set worrying precedents for the usage of digital content, potentially normalizing unauthorized access and modification of creative works. These concerns are echoed by numerous stakeholders, including over 400 Hollywood professionals who have vocally opposed the notion, arguing that it undermines the efforts and economic stability of creative industries. Such an exception not only jeopardizes the income and integrity of artists and writers, but also poses a risk of diminishing the quality and diversity of content available, as creators may lose the incentive to produce original works if their livelihood is threatened by such sweeping exceptions to copyright [source](https://www.indiewire.com/news/business/ai‑open‑letter‑openai‑google‑copyright‑exemptions‑1235104394/).

Adding to the complexity are the ethical dilemmas posed by the potential misuse of AI models trained on such comprehensive yet uncompensated datasets. The lack of clear guidelines and the potential for biased, unfair, or unaccountable AI decision‑making raises alarms about the ethical deployment of these technologies. As these AI systems are deployed into public domains, their influences on cultural norms and values could be both profound and unpredictable. The challenge lies in ensuring these powerful tools do not inadvertently perpetuate inequality or infringe upon personal liberties. In this evolving landscape, the need for transparent regulation and ethical oversight becomes paramount to balance technological growth with societal well‑being [source](https://techpolicy.press/generative‑ai‑and‑copyright‑issues‑globally‑ani‑media‑v‑openai).

The request by OpenAI and Google to train AI models on copyrighted material without obtaining licenses not only raises significant legal and ethical questions but also presents substantial political consequences. The outcome of this request will fundamentally influence the interplay between government policies and the burgeoning field of artificial intelligence. Politically, granting the exemption may be perceived as the government prioritizing technological advancement over the protection of intellectual property rights, potentially riling creators and intellectual property advocates. Such a decision could lead to heightened demands for robust AI regulations and oversight to ensure that AI development does not trample existing copyrights and the interests of creative industries. More broadly, the policy stance on this matter will reflect the government's position on innovation versus tradition—a pivotal point in shaping the future of AI governance.

Conversely, denying the request could position the government as a staunch defender of intellectual property, aligning with the current regulatory frameworks and potentially preserving the sanctity of creative works. This may bolster the creative industry’s confidence in the legal system's ability to protect their creations against potential misuse by large technology firms. However, it might also trigger concerns about stifling innovation, as tech firms may argue that the decision thwarts progress essential for maintaining competitiveness on a global scale, especially against countries like China where less stringent data usage laws might prevail. Hence, political leaders will face pressure to find a middle ground that promotes technological progress while genuinely respecting intellectual property rights.

The international dimension of this political impact should not be underestimated. As AI development becomes a critical aspect of national competence and security, the way the U.S. handles fair use laws for AI training could have ripple effects on its international competitiveness. Countries that strike a more lenient balance could potentially lure companies seeking fewer restrictions, thereby shifting technological leadership overseas. Therefore, the U.S. government's decision on this matter will likely weigh heavily between supporting domestic innovation and preserving a global competitive edge, akin to a balancing act of legislative foresight and strategic geopolitical considerations.

The request by OpenAI and Google to train AI models on copyrighted content without licenses is a pivotal moment in the advancement of artificial intelligence. This proposal could drastically alter the landscape of AI innovation and international competitiveness. By removing the constraints imposed by copyright laws, AI models can be developed on more diverse and comprehensive datasets, potentially accelerating technological growth and maintaining U.S. leadership in the AI field. Proponents argue that this approach is not just about fostering innovation but also about addressing national security concerns, as cutting‑edge AI capabilities are essential for economic and strategic advantages in the global arena. However, this bold step poses significant challenges and raises questions about the balance between innovation and intellectual property rights. [Nieman Lab](https://www.niemanlab.org/reading/openai‑and‑google‑ask‑the‑government‑to‑let‑them‑train‑ai‑on‑content‑they‑dont‑own/)

Critics of the proposed copyright exception argue that it could undermine traditional creative industries, potentially devaluing content creators' rights and revenues. With AI models utilizing copyrighted material without permission, the fear is that it might set a precedent for the large‑scale appropriation of creative works, thereby infringing upon intellectual property laws and ethical business practices. Hollywood and other creative sectors have been vocal about their opposition, emphasizing the economic significance and cultural contributions provided by these industries. They argue for a more structured approach where AI companies negotiate licenses, thereby ensuring a fair exchange of value and continued investment in creative content. [Hollywood Reporter](https://www.hollywoodreporter.com/business/business‑news/hollywood‑pushes‑back‑openai‑google‑argument‑copyright‑1236166626/)

The international race for AI supremacy adds another layer of complexity to this debate. As countries like China aggressively develop their AI technologies, U.S. firms feel the pressure to maintain a competitive edge. Without the freedom to use vast amounts of data, there's a real risk of falling behind in innovation. Granting the copyright exception for AI training could be seen as a mechanism to ensure U.S. firms remain at the forefront, allowing them access to critical data that can enhance model accuracy and capability. However, this could also lead to tensions on the international stage, particularly if it results in an uneven playing field or sets precedents that other countries, with different copyright laws, might not follow. This scenario underscores the need for careful international dialogue and policy alignment to navigate these competitive and ethical concerns effectively. [Forbes](https://www.forbes.com/sites/virginieberger/2025/03/15/the‑ai‑copyright‑battle‑why‑openai‑and‑google‑are‑pushing‑for‑fair‑use/)

In navigating the intersection between artificial intelligence development and copyright laws, a nuanced approach is essential. OpenAI and Google's requests for copyright exceptions underscore the complexities involved in balancing technological advancement with the protection of creative industries. The ongoing legal debates illustrate the tension between ensuring AI models are trained on comprehensive data sets and respecting the intellectual property rights of creators. As the U.S. Copyright Office deliberates on this critical issue, the outcome could set a precedent that impacts not only the AI industry but also the broader landscape of copyright law and creative works [1](https://www.niemanlab.org/reading/openai‑and‑google‑ask‑the‑government‑to‑let‑them‑train‑ai‑on‑content‑they‑dont‑own/).

It is crucial to recognize the multifaceted implications of either granting or denying the requested copyright exceptions. On one hand, granting these exceptions could accelerate AI innovation and maintain the U.S.'s competitive edge in the global tech economy. On the other hand, it could undermine the economic viability of creative sectors that rely heavily on the protection of copyrighted material. The creative industry, as highlighted by the strong opposition from Hollywood, fears the potential devaluation and exploitation of creative work without appropriate compensation [7](https://www.indiewire.com/news/business/ai‑open‑letter‑openai‑google‑copyright‑exemptions‑1235104394/)[12](https://www.hollywoodreporter.com/business/business‑news/hollywood‑pushes‑back‑openai‑google‑argument‑copyright‑1236166626/).

Ultimately, a balanced approach might be the most prudent path forward, one that considers the interests of both AI developers and copyright holders. Collaborative solutions involving licensing agreements or the creation of alternative datasets could be explored to satisfy both parties. This approach not only fosters innovation but also safeguards the creative economy and its contributions to cultural heritage. As experts emphasize, current copyright laws may need to evolve to meet the unique challenges posed by modern AI technologies, ensuring that innovation can flourish alongside a robust creative industry [9](https://www.pymnts.com/news/artificial‑intelligence/2025/ai‑regulations‑tech‑giants‑and‑hollywood‑converge‑on‑white‑house‑ai‑strategy/).

Moving forward, the dialogue between AI innovators and copyright advocates is crucial in shaping a future that embraces technological progress while honoring the integrity and economic significance of creative rights. This discourse needs to incorporate input from a diverse array of stakeholders, including technology companies, legal experts, content creators, and governmental bodies. By working towards a shared understanding of "fair use" in the context of AI, the stakeholders can better align on policies and practices that balance innovation with respect for intellectual property [6](https://www.forbes.com/sites/virginieberger/2025/03/15/the‑ai‑copyright‑battle‑why‑openai‑and‑google‑are‑pushing‑for‑fair‑use/).