Big photo: With at least one court verdict that AI training is not as perfect use, Openai Trump is looking at the upcoming AI action plan to help resolve the ongoing copyright disputes. In July, the plan can potentially classify AI training as proper use, providing AI companies unrestricted access to important training data. Openai argues that such a step is necessary for the US to maintain its competitive lead in the AI race against China.
American courts are currently struggling with whether AI training is eligible as fair use, the right holders argued that AI models potentially pose a market risk by diluting their place and creative output. Openai and many other AI companies are entangled in many cases, saying that AI changes copyright tasks rather than serving as a direct option. However, a landmark ruling has already favored the rights holders, it has been determined with a judge that AI training does not make proper use, citing Thomson Reuters’ Legal Research firm, cited his direct threat to Westlav.
This debate is also coming up internationally, as the countries want to balance copyright security with the increasing demand for AI training.
Openai claims that its models do not repeat copyright tasks for public consumption, but take out patterns and insights to generate new and specific materials. The company argues that this approach aligns with the basic principles of copyright and fair use theory.
“OpenaiI’s model is trained not to work for consumption by the public. Instead, they remove the tasks and extract patterns, linguistic structures and relevant insights,” Openai explained. “This means that our AI model training aligns with the main objectives and fair use theory of the training copyright, using existing tasks to create completely new and different without eradicating the business value of those current tasks.”
During a public comment period, OpenAII urged the US to move its copyright strategy to promote the “freedom of learning” of the AI industry, warning that American companies were banned from reaching copyright data – while there is no limit to Chinese firms – which can give the US the cost of its AI leadership.
“The federal government can secure both the freedom of Americans to learn from AI and can avoid seizing our AI lead to PRC by preserving the ability of the US AI model to learn from copyright content,” Openai said.
The company also asked for legal safety for AI firms, citing stress caused by a patchwork of state rules. By 2025, the Vidhan Tracker Multistate has identified 832 AI-related laws. Openai warned that reflecting the European Union strict regulatory approaches can prevent innovation, apply cumbersome compliance costs, and weaken economic competition and national security.
Instead, it proposed a federal law that predetermines the rules of the state, offering a voluntary public-private partnership where AI companies share the knowledge of the industry in exchange for liability security.
Openai urged the US to lead a global discussion on copyright and AI to prevent less new countries from implementing restrictive legal framework on American firms. This involves assessing data availability and ensuring that American companies maintain important training inputs. With China’s Rapid AI progression, such as the open-sourced deepsek model, Openi warned that the American lead in AI is narrow and requires immediate action to maintain it.
Additionally, the company insisted that national security depends on access to AI training data and has been called for a balanced approach, protecting intellectual property rights. As the US Copyright Office prepares further guidance on AI training in its upcoming report, bets are high for both AI firms and copyright holders.
