WASHINGTON, DC - JANUARY 21: OpenAI CEO Sam Altman (R), accompanied by U.S. President Donald Trump, ... [+] speaks during a news conference in the Roosevelt Room of the White House on January 21, 2025 in Washington, DC. Trump announced an investment in artificial intelligence (AI) infrastructure and took questions on a range of topics including his presidential pardons of Jan. 6 defendants, the war in Ukraine, cryptocurrencies and other topics. (Photo by Andrew Harnik/Getty Images) Getty Images Artificial intelligence powerhouses OpenAI and Google are aggressively lobbying the U.S. government to classify AI training on copyrighted data as "fair use." Their objective, framed as a matter of national security, is positioned to secure a competitive advantage against international rivals, particularly China. However, this proposal raises profound legal, ethical, and economic questions, illustrated sharply by recent high-profile cases involving companies like Meta and recent lawsuits by French publishers. OpenAI’s Strategic Lobbying And The 'Intelligence Age' OpenAI and Google recently submitted extensive policy proposals in response to a request for public comments from the White House Office of Science and Technology Policy. The submission is part of the U.S. government's broader Artificial Intelligence Action Plan, initiated under an executive order from the Trump administration. Their proposals argue that limiting AI training on copyrighted materials could weaken America's technological edge and slow down innovation, positioning national security as a core justification for broad fair use protections. Sam Altman, CEO of OpenAI, labels the current period as an "Intelligence Age," suggesting restrictive copyright laws could inadvertently empower geopolitical rivals such as China. He argues that U.S. leadership in AI directly correlates with national security, economic prosperity, and democratic ideals. Google reinforces this narrative, emphasizing how current copyright restrictions are excessively cautious, inspired by stringent European models. Google's stance is that fair use and text-and-data mining exceptions are vital, asserting that restrictions create unnecessary complexity, unpredictability, and delays, ultimately hampering American technological innovation. Both companies present a clear narrative: if U.S. companies are unable to freely train AI systems, American technological leadership could be lost, primarily to China, where companies operate with fewer regulatory constraints. The Meta Scandal: A Warning on the Limits of Fair Use The implications of broadening the interpretation of fair use are starkly highlighted in the recent Meta controversy. Meta faced accusations of torrenting copyrighted books, without permission, to train AI models, leading authors to launch a landmark lawsuit. Documents revealed deliberate concealment strategies by Meta, including using Amazon Web Services to mask their actions. MORE FOR YOU Google Confirms Play Store App Deletion—What You Do Now ‘Reacher’ Dethroned In Amazon Prime Video’s Top 10 List By A New Show FBI Warning—Gmail, Outlook And VPN Users Need To Act Now Authors sued Meta for copyright infringement, asserting these actions constituted outright piracy, not fair use. The authors' lawsuit introduced a particularly damning argument, invoking the "Bob Dylan defense," ironically referencing lyrics illustrating inequitable treatment, where large corporations seemingly skirt laws with impunity. Additionally, French publishers, led by the National Publishing Union, the National Union of Authors and Composers, and the Society of Men of Letters, have taken legal action against Meta. These organizations, which defend the interests of authors and publishers, filed a complaint in a Paris court alleging that Meta engaged in systematic copyright infringement and economic "parasitism" by using copyrighted works to train its AI models. This case highlights growing global resistance to the unchecked use of creative works in AI training and sets a precedent for future legal battles beyond the United States. The Reality of AI Training: Data, Not the Model, Drives Value AI companies often claim that their models do not replicate copyrighted works but instead "learn" from them by extracting patterns, linguistic structures, and contextual insights. However, the core function of generative models is to statistically mimic the system that produced the training data. This means that a language model can be prompted to write in the style of a specific author or mimic an artist’s work because it has encoded patterns from the original material. From a technical perspective, machine learning is not "learning" in the way a human does, it is primarily a large-scale data compression mechanism. During training, AI models encode statistical patterns of their datasets, retaining a significant portion of those patterns even after fine-tuning. This makes it possible for models to regenerate training data within reasonable margins of error, effectively reproducing copyrighted works in ways that require attribution or licensing. Critics like Chomba Bupe, tech entrepreneur and expert in Machine Intelligence, argue that this undermines claims of fair use because AI models do not create truly novel content but rather recombine compressed versions of copyrighted materials. This strengthens the argument that AI companies should compensate or seek explicit consent from content creators whose works are included in training datasets. Furthermore, the ongoing lawsuits against AI firms could serve as a necessary correction to push the industry toward genuinely intelligent machine learning models instead of data-compression-based generators masquerading as intelligence. If legal challenges force AI firms to rethink their reliance on copyrighted content, it could spur innovation toward creating more advanced, ethically sourced AI systems. Fair Use Doctrine: Legal Uncertainty and Industry Tensions Central to OpenAI and Google's argument is the legal doctrine of fair use, historically permitting limited transformative uses of copyrighted materials. AI companies assert their algorithms don't reproduce copyrighted works directly for public consumption but instead analyze patterns, context, and structures, creating transformative outputs. Yet, recent landmark court decisions cast doubt on this interpretation. Notably, the Thomson-Reuters and Westlaw case demonstrated AI-generated outputs could significantly undermine established markets, not merely complement or enhance them. Moreover, OpenAI faces multiple significant lawsuits, including from major publishers like The New York Times, highlighting ongoing contention around fair use applicability in the AI era. Relying only on fair use as a legal shield is a precarious business model. If your business model relies on obtaining raw materials for free, materials that are likely protected by copyright, then you are assuming a liability from the outset. Investors may see this legal risk as a structural flaw, especially given the increasing number of lawsuits filed against AI firms. Evaluating National Security Claims: Real Risk Or Regulatory Loophole? Both OpenAI and Google emphasize national security concerns, warning that overly restrictive copyright laws could allow China to surpass U.S. technological capabilities. They frequently cite China's rapid AI advancements, exemplified by DeepSeek AI, which recently drew attention from Chinese President Xi Jinping. However, it can be argued that national security justifications risk becoming a convenient regulatory loophole. The invocation of geopolitical risks might serve as leverage to grant AI companies overly broad privileges, potentially undermining intellectual property protections and creators' rights. Recommendations: Finding a Sustainable Balance A sustainable solution must reconcile technological innovation with creators' economic interests. Policymakers should develop clear federal standards specifying fair use parameters for AI training, considering solutions such as: Licensing and Royalties: Transparent licensing arrangements compensating creators whose work is integral to AI datasets. Curated Datasets: Government or industry-managed datasets explicitly approved for AI training, ensuring fair compensation. Regulated Exceptions: Clear legal definitions distinguishing transformative use in AI training contexts. These nuanced policies could encourage innovation without sacrificing creators’ rights. The lobbying by OpenAI and Google reveals broader tensions between rapid technological growth and ethical accountability. While national security concerns warrant careful consideration, they must not justify irresponsible regulation or ethical compromises. A balanced approach, preserving innovation, protecting creators’ rights, and ensuring sustainable and ethical AI development, is critical for future global competitiveness and societal fairness Follow me on Twitter or LinkedIn. Editorial StandardsForbes Accolades