Confirm training on lawfully accessed works as fair use / TDM exception

“We believe that, in general, training generative AI constitutes fair use under current U.S. law. Using creative works to train generative AI fits with the long line of cases that has found that non-consumptive, technological uses of creative works in ways that are unrelated to the expressive content of those works are transformative fair uses”

“We are clear in our view that the training of AI models is a non-expressive use of open web content, which boosts economic growth, fosters scientific advancement, and enables the creation of valuable new works.”

“techUK believes that Option 2 – the broad TDM exception – would provide the strongest competitive advantage for the UK's AI sector and create significant opportunities for economic growth and will be the preferred option of many.”

“The federal government can both secure Americans’ freedom to learn from AI, and avoid forfeiting our AI lead to the PRC by preserving American AI models’ ability to learn from copyrighted material.”

“And the federal government should embrace policy frameworks that preserve access to data for fair learning, advance a risk-based approach to AI applications based on existing regulations, and preempt a chaotic patchwork of state-level rules on frontier AI development.”

Same sentence carries both the fair-learning and preemption asks; recorded once per ask.

“In addition, we urge the administration to clarify that existing copyright law protects the ability of developers to train models. Getting this right is critical to ensuring American competitiveness in AI.”

“We believe a broad text and data mining exception, as described in Option 2, is the most effective way to achieve the Government’s objectives to unlock and leverage the benefits of AI in the UK.”

Submitted via IPO survey 25 Feb 2025; summary published 2 Apr 2025 at the Commons SITC's request.

“These cases demonstrate that the fair use defense enables courts to consider all the fair use factors and apply them in the context of specific facts. MPA submits that this type of inquiry is the appropriate way to deal with the many types of potential infringements that may arise under the broad umbrella of “training” a generative AI system.”

Position is case-by-case fair-use analysis, opposing categorical treatment of training as fair use; also opposes new AI-specific copyright legislation.

“Retaining the UK’s gold standard copyright protections - and ensuring the law is enforceable and respected in the face of the challenges posed by GAI - will create incentives for GAI developers to enter into licence agreements with rights holders, ensuring a steady flow of quality, human-authored works for GAI training.”

“The AI Action Plan should encourage the continued development of free market licensing to support a symbiotic relationship between content creators, publishers, and AI developers.”

“Specifically, the nation’s AI policy and AI Action Plan should respect intellectual property, support the development of voluntary licensing in free markets, and promote appropriate transparency and fair competition in the AI and technology sector.”

The filed RFI comments; doc-nma-actionplan-2025 is the accompanying press release.

“The report presents thoughtful and balanced analysis of fair use in the context of AI training, emphasizing that AI training is not categorically fair use and that whether a use qualifies as a fair use is a matter of context and degree.”

“In the report, the Copyright Office takes the sensible and correct approach that AI training is not categorically fair use and that whether a use qualifies as a fair use is a matter of context and degree.”

“The Secretary of State must now definitively rule out any new copyright exception, bringing an end to the uncertainty created by this prolonged process over the past year. This will send a clear message to AI developers that they must enter into licensing agreements with the UK’s media and creative copyright owners”