How Media Organizations Are Navigating Partnerships with Large Language Model Providers and the Rise of Generative Engine Optimization
The relationship between media organizations and artificial intelligence companies has undergone a seismic shift in just 18 months. What began as a confrontational legal battle has transformed into a complex ecosystem of licensing agreements, revenue-sharing arrangements, and strategic partnerships. This evolution reveals a fundamental truth about the AI era: publishers cannot afford to exclude themselves from the systems determining how news reaches audiences, even if that means sharing content with potential competitors.
The inflection point arrived in late 2023, when The New York Times filed a groundbreaking copyright lawsuit against OpenAI and Microsoft, asserting that millions of its articles were used without permission to train ChatGPT.[1] Rather than merely defending the lawsuit, however, publishers soon realized that legal opposition alone would not determine their fate in an AI-driven information ecosystem. If the technology companies would license content regardless, publishers had to ensure they received financial compensation and attribution.
By 2024, this strategic realization crystallized into an unprecedented wave of licensing deals. The Associated Press, Financial Times, Reuters, News Corp, The Washington Post, and dozens of smaller publishers all negotiated agreements to license their content to OpenAI, Google, Meta, Amazon, and other AI firms. These deals signal a profound acceptance: publishers are not fighting AI adoption; they are negotiating their role within it. The question has shifted from "Should AI companies have access to our content?" to "What is fair compensation, and how do we retain traffic and influence?"
What makes this moment significant for communications professionals is that it demonstrates how quickly media businesses can shift from adversarial to cooperative relationships with technology companies when the alternative is obsolescence. A comms strategist advising a media organization in 2024 had to navigate a fundamentally different landscape than one advising in 2022. The question was no longer whether to engage with AI but how to structure that engagement to benefit the organization's bottom line and long-term viability.
The financial scale of media-AI partnerships has expanded dramatically, with total compensation now reaching into the billions of dollars. These deals fall into two primary categories: one-time training data licensing, and ongoing visibility and attribution agreements.
OpenAI has emerged as the most aggressive acquirer of publisher content, signing deals with nearly every major news organization. The AP was first, announcing in July 2023 a two-year arrangement granting OpenAI access to its news archive dating back to 1985 and providing ongoing content licensing.[2]
In December 2023, OpenAI inked a landmark partnership with Axel Springer, the German media conglomerate. Details emerged that the deal was worth approximately $25 million over multiple years, with a one-off payment for historical training data and variable back-end fees tied to product usage.[3] This structure became a template: OpenAI would pay for access to training data and additional compensation as it commercialized products using that content.
By 2024-2025, OpenAI had signed agreements with the Financial Times[4], News Corp (owner of the Wall Street Journal and The Times of London)[5], Condé Nast[6], Time magazine[7], and The Washington Post[8]. Each deal offered unique terms, but the pattern was consistent: OpenAI gained access to premium journalism for both training and display purposes, while publishers secured revenue and potential traffic from ChatGPT users clicking through to read full articles.
Meta Platforms and Google have also entered the licensing market aggressively. In March 2026, News Corp announced a multi-year deal with Meta worth up to $50 million annually, positioning News Corp as a primary content provider for Meta's AI products.[9] This represents the largest single annual licensing arrangement to date.
Google's approach has been more selective but no less strategic. Google signed a deal with News Corp valued at $5-6 million annually for new AI content development[10], and the Associated Press announced a licensing partnership with Google as well.[11]
These partnerships create an asymmetrical incentive structure. AI companies receive access to premium journalism at a relatively low cost compared to their total revenue, while publishers secure a new revenue stream but remain dependent on the growth of these AI platforms for impact and reach.
Despite the wave of licensing deals and the promise of AI-driven traffic, publishers face a troubling reality: overall traffic from search and AI sources has collapsed. This paradox defines the current media crisis and fundamentally shapes how communications leaders must advise their organizations.
In May 2024, Google introduced AI Overviews, an AI-generated summary feature within search results that directly answers user queries without requiring users to click through to publisher websites. The feature was positioned as an enhancement to search, providing faster access to information. For publishers, it became an existential threat.
Data from multiple research firms documents the impact with startling clarity. Chartbeat reported that global organic search traffic to publishers dropped by 33% from November 2024 to November 2025.[12] Other research suggests even steeper declines: one analysis found page views from Google Search referrals fell 34-46% depending on publisher type and market.[13] For smaller publishers and niche news outlets, the decline exceeded 50%.[14]
NPR documented publisher executives describing the situation as an "extinction-level event,"[15] a phrase that captures the existential anxiety driving strategy decisions at media companies.
Google AI Overviews represent what researchers call "zero-click search"—answers delivered directly to the user without requiring navigation to source websites. While Google does include attributed links to source articles in some AI Overviews, the design prioritizes the summary. Users frequently get the information they need without ever seeing the publisher's website, never encounter advertising, and generate no page views or advertising revenue for the publisher.
ChatGPT and Perplexity AI have demonstrated a contrasting model: both platforms deliver citations that link directly to source articles, potentially driving traffic. However, the referral value pales in comparison to losses from Google. TollBit research found that AI search engines send 96% less referral traffic to news sites than traditional Google search.[16] Even though ChatGPT referrals grew 25x from under 1 million in early 2024 to over 25 million in 2025, this gain represents only a fraction of the traffic lost to Google AI Overviews.[17]
This creates an uncomfortable reality for publishers and the communications leaders advising them. Content licensing deals with AI companies—while important for immediate revenue—cannot mathematically offset the structural loss of traffic from AI-driven search. A publisher that receives $10 million annually in licensing fees but loses 40% of its search-driven traffic (often worth far more in advertising and subscription revenue) has not solved its problem; it has merely slowed the rate of decline.
The implication for communications strategy is profound: traditional arguments about media's value (reach, influence, monetization through ads) must be rebuilt for an AI-mediated information environment. Publishers cannot simply wait for the market to stabilize; they must actively reshape their content strategy, audience relationship, and revenue model to thrive in a world where search is increasingly mediated by large language models.
In response to the traffic crisis and the rise of AI search, a new discipline has emerged: Generative Engine Optimization (GEO). This represents perhaps the most significant evolution in content strategy since the rise of Search Engine Optimization (SEO) in the 1990s, and it carries profound implications for how communications professionals must think about media relations, content creation, and audience engagement.
GEO acknowledges a fundamental shift in how audiences discover and consume information. Rather than typing a query into Google and scanning a list of blue links, users increasingly type a question into ChatGPT, Perplexity, or Google's AI Overview and receive a natural-language answer synthesized from multiple sources. The question for communications leaders is: Will your organization's content be included in that answer? Will it be attributed? Will the summary be accurate?
Research from digital marketing agencies and SEO firms has identified key GEO principles:[18]
Clarity and Specificity in Headers and Metadata: AI models rely on clear structural signals when extracting and summarizing content. Articles with well-organized headers, clear topic sentences, and descriptive metadata are more likely to be selected as sources.
Direct Answers to Common Questions: Rather than burying key information in prose, high-performing GEO content structures answers to the specific questions AI systems are designed to answer. If users ask "What is generative AI?" a GEO-optimized article will provide a direct, one-sentence definition high in the content.
Primary Research and Unique Data: AI systems are trained to identify original research and data. Organizations that conduct surveys, experiments, or analysis generate content that AI systems are more likely to cite as authoritative sources.
Topic Authority Rather Than Keyword Density: Traditional SEO optimized for keyword repetition. GEO optimizes for demonstrating depth of knowledge on a topic. An article that comprehensively covers a topic from multiple angles is more likely to be selected than one that repetitively uses a single keyword.
Technical Optimization for AI Crawling: Structured data, schema markup, and clear content hierarchy make it easier for AI systems to understand and extract information from web pages.
For public relations and communications professionals, GEO opens a new frontier. The traditional PR goal of "getting coverage" in major outlets remains valuable, but it now competes with a different objective: ensuring that your organization's information is selected, accurately represented, and properly attributed when AI systems synthesize answers.
A healthcare company might pitch a story to journalists about a new clinical study. But GEO strategy also means optimizing the original study's publication (white paper, research page, or announcement) so that ChatGPT and other AI systems can easily find, understand, and cite it when answering health-related questions. The two strategies are complementary but require different execution.
While publishers and AI companies negotiate licensing deals, the legal status of AI training on copyrighted material remains contested. Multiple lawsuits are working through the court system, with significant implications for the entire ecosystem.
The New York Times' lawsuit against OpenAI and Microsoft remains the marquee case. Filed in December 2023, the suit alleged that OpenAI used millions of Times articles without permission to train ChatGPT and that ChatGPT now competes with the Times' own business.[19]
As of early 2026, the case continues through discovery and motion phases, with no final ruling. However, updates suggest that the central issue—whether ChatGPT sometimes "regurgitates" Times articles verbatim—has become critical to the outcome.[20] If the court finds that ChatGPT reproduces substantial portions of copyrighted articles without paraphrasing, OpenAI's fair use defense becomes weaker.
Beyond the Times case, 2024-2025 saw multiple federal court decisions on whether AI training constitutes fair use. The landscape is mixed but trending toward acceptance of AI training under specific conditions.
In several district court decisions, judges found that using copyrighted works to train AI models—particularly when the training is transformative and the primary market of the original work is not harmed—can constitute fair use.[21] This principle comes from established fair use doctrine (transformative use, limited market harm), but its application to AI is novel.
However, courts have also signaled that certain practices—such as ingesting copyrighted datasets wholesale without transforming them—may not be fair use.[22] The distinction matters: fair use for training on lawfully obtained data may not extend to datasets compiled specifically for training AI systems in violation of copyright.
The Copyright Office has also engaged in this debate, publishing guidance on AI and copyright that signals the U.S. government's position that fair use protections should apply to transformative AI uses, but with important caveats around harm to original markets.[23]
For technology companies, licensing deals with publishers serve a dual purpose. They provide revenue to publishers and, critically, they reduce legal exposure. If OpenAI and Meta can demonstrate that they obtained consent from the largest publishers to use their content, the fair use defense in litigation becomes stronger: the company can argue that licensing represents industry best practice and that disputes with other publishers are business disagreements, not attempts to evade copyright law.
This dynamic explains why AI companies have aggressively pursued licensing deals even when not legally required. The deals are simultaneously business strategy and legal strategy.
The transformation of media-AI relationships carries distinct implications for communications professionals advising media organizations, technology companies, brands, and public figures.
Publisher leadership must recognize that licensing deals are revenue supplementation, not salvation. The financial gains from licensing (millions of dollars annually) cannot offset the structural losses from AI-driven search (often tens of millions in lost traffic-driven revenue). Media companies must simultaneously pursue licensing agreements while rebuilding business models less dependent on search traffic.
This might include accelerated investment in direct relationships with audiences (email newsletters, subscriptions, membership programs), diversification of revenue streams (events, education, data products), and content strategies optimized for both human readers and AI systems. Communications leaders advising publishers should help articulate this dual focus to stakeholders.
AI company communicators must manage a delicate balance. Publishing licensing deals signal commitment to supporting journalism and respecting intellectual property. But these deals also risk criticism that the company is enriching a small number of major publishers while marginalizing smaller outlets that lack bargaining power.
A transparent communications strategy that addresses how the company supports creators and publishers across scale—through both paid licensing and technical features that drive attribution and traffic—is essential for maintaining trust.
The rise of GEO creates new opportunities for brands to ensure accurate representation in AI-generated answers. Marketing and communications teams should develop GEO strategies alongside traditional PR and content marketing. This includes optimizing owned-media properties (websites, reports, data) for AI discoverability and partnering with publishers on GEO-aligned content strategy.
The media-AI landscape remains in flux. Google may adjust AI Overviews in response to publisher pressure. Licensing deal terms may shift as AI companies' market positions change. Copyright litigation will ultimately settle certain legal questions. Communications leaders must treat the current moment not as a final equilibrium but as a waypoint in ongoing evolution.
The core strategic lesson is this: AI systems are now the primary gatekeepers of information for growing portions of the audience. Organizations that succeed in an AI-mediated information environment will be those that actively engage with these systems—through licensing, through GEO-optimized content, through strategic partnerships, and through clear communication about their value and trustworthiness to both humans and machines.
Exclusive workshops, speaking engagements, and strategic consulting for senior communications and PR professionals navigating the AI era.