Since the introduction of the AI Overview, search and publishing experts have been paying close attention to how AI companies should treat the content used to train their models. Google has now shared its status. It emphasizes fair use and offers opt-out options, while highlighting paid agreements in certain situations.
In a policy paper published on June 25, Google shared that training models on publicly available web data are considered “transformative, non-disclosing uses” that should remain protected under fair use in the U.S. The company highlights exit controls and existing copyright law as their main solutions to address publishers’ concerns.
The paper, “A Pragmatic Approach to AI Governance in America,” pulls together points that Google has previously shared. It comes at a time when regulators and publishers want more, they don’t just want an exit but also a clear explanation and sometimes even compensation. For publishers figuring out how to manage AI access to their content, it provides a useful insight into where Google stands.
Copyright position of Google
Google likens AI training to “an art student who gets inspiration from a gallery walk.” It also suggests that the same level of protection should be extended internationally with the exception of text and data mining.
For site owners who do not want their content manipulated, Google recommends using machine-readable controls like Google-Extended in their robots.txt. When AI results copy an existing work, the solution is not to filter to judge that the output is “very similar,” but to rely on known processes of recognition and reduction, as stated in the paper.
Google is also looking at new ways to create value, such as partnerships with websites that provide content that helps keep AI responses up-to-date and accurate, and agreements to pay for access to exclusive, non-public content. The paper does not state any plans, goals, or timelines.
Where the Position Lives
This month, the UK’s CMA introduced a new ethical requirement that gives websites the option to opt out of AI search features and requires Google to cite publisher content. The regulator stated that this move is intended to help increase the bargaining power of publishers. Google has already started testing the opt-out toggle, although the reports available to publishers to help them decide don’t yet include click data.
American publishers make their position even clearer. Next Digital Content recently sent a cease-and-desist letter to the Common Crawl Foundation, insisting that “copyright law is not an exit law.” This means that scrapers must seek permission before using content, rather than publishers asking for takedowns. This idea directly challenges the exit model discussed in the Google paper.
Why This Matters
This paper highlights Google’s position as policymakers consider new regulations. Google recommends keeping its current approach unchanged.
Publishers and regulators want more than what paper currently offers. They ask for compensation, pre-scratching permission, and detailed click-through rate data. In response, this paper provides controls and manages conversations for each individual.
Looking Forward
These are policy positions, not product commitments. The basic partnerships and content deals Google is talking about could affect how much value reaches publishers, but the paper leaves the details vague. Be aware of whether Google links programs, terms, or statistics to the exchange rate language it includes in its policy documents.
Featured Image: FotoField/Shutterstock
