The Great AI Copyright War: A Global Tally of Lawsuits and Licensing Deals

The rapid ascendancy of generative artificial intelligence has triggered a profound transformation in the global media landscape, pitting legacy news publishers against the Silicon Valley giants that have built their multi-billion dollar models on the back of human-authored content. As AI companies race to refine tools like ChatGPT, Claude, and Perplexity, the news industry finds itself at a historic crossroads: engage in high-stakes litigation to defend intellectual property rights or secure lucrative, albeit controversial, licensing deals to ensure financial survival.

This battle has evolved into a global conflict. From the United States and Canada to Brazil, India, and Japan, publishers are asserting that the unlicensed "scraping" of their reporting, analysis, and images constitutes the greatest existential threat to the Fourth Estate since the advent of the internet.

The Frontlines: A Growing Wave of Litigation

The most recent major escalation occurred in October 2026, when USA Today Co became the latest heavyweight to join the fray. The publisher, which oversees 19 major publications including The Tennessean, Indy Star, and the Detroit Free Press, filed a lawsuit against OpenAI seeking $250 million in damages. The complaint alleges that OpenAI’s unauthorized use of hundreds of thousands of articles has caused "real and continuing damage" to its brands, effectively cannibalizing their subscription and advertising revenue by repackaging their work for AI users.

This legal action is part of a broader, systemic trend. In September 2026, The Seattle Times and Newsday launched a joint lawsuit against both OpenAI and Microsoft. The publishers argued that the "transformative" fair use defense—the primary shield used by AI companies—is fundamentally flawed. They contend that there is nothing transformative about copying high-quality, expensive journalism without compensation to create products that compete directly with the original publishers for the exact same readers and advertisers.

The conflict has also moved beyond US borders. In August 2026, the Argentine publisher Editorial Perfil sued OpenAI and Microsoft, becoming the first Spanish-language news organization to take such action. Similarly, Japanese giant Yomiuri Shimbun filed a lawsuit against Perplexity, seeking roughly £9.9 million in damages for the alleged unauthorized acquisition of over 119,000 articles.

The Perplexity Controversy

Perplexity AI, in particular, has become a lightning rod for criticism. Unlike traditional chatbots that generate text from static training data, Perplexity’s "answer engine" uses real-time retrieval-augmented generation (RAG) to scan the web and synthesize results. Major organizations, including The New York Times, CNN, and The Chicago Tribune, have initiated litigation against the firm, arguing that its process of "stealing" content from behind paywalls—content they spent millions to produce—is a direct violation of copyright law and a form of industrial-scale unfair competition.

The "Who’s Who" of AI Licensing Deals

While many publishers have chosen the path of war, others have opted for diplomacy, signing multi-year licensing agreements. These deals are designed to turn an existential threat into a revenue stream, though they remain controversial within the industry.

• News Corp and Meta: In a massive move, News Corp secured a multi-year deal with Meta worth up to $50 million annually, allowing the social media giant to use content from titles like The Wall Street Journal and The Times (UK) to power its AI assistants.
• OpenAI Partnerships: OpenAI has been aggressively signing deals to stave off litigation. In September 2026, it finalized agreements with India’s two largest newspaper groups, Bennett, Coleman and Co (publisher of The Times of India) and The Indian Express Group, potentially worth $5 million and $3 million per year, respectively.
• Usage-Based Models: Some publishers, such as the UK’s Reach and the US-based People Inc, have favored usage-based licensing. This model, which compensates publishers based on how often their content is accessed by an AI, is seen by some as a more sustainable, transparent approach to the value exchange between AI companies and the media.

The Legal and Ethical Implications

The core legal question remains: Does training an AI model on copyrighted data constitute "fair use"?

AI firms argue that their models are "grounded" in publicly available data and serve the public interest by facilitating innovation, medical research, and enhanced access to information. Microsoft and OpenAI consistently emphasize their willingness to negotiate, often citing their partnerships with organizations like The Lenfest Institute as evidence of their commitment to supporting local journalism.

However, the publishers’ counter-argument is stark: The AI boom, as currently structured, represents a "death knell" for the business models that fund professional, fact-checked reporting. When an AI provides a user with a "verbatim or near-verbatim" summary of an article, the user has no incentive to click through to the original source. This destroys traffic, evaporates ad revenue, and undermines the subscription model, leaving the media ecosystem starved of the resources needed to investigate the very facts the AI models are consuming.

Furthermore, the issue of "hallucinations"—where AI models invent facts and falsely attribute them to reputable publishers—has added a layer of reputational risk to the copyright dispute. As The Chicago Tribune and other outlets have documented, the misattribution of bogus medical or political advice to trusted news brands threatens the very credibility that journalism relies upon.

Official Responses and Future Outlook

The landscape is volatile, with the courts likely to spend the next several years determining the boundaries of copyright in the age of generative AI.

• The OpenAI Stance: OpenAI continues to maintain that its models are trained on publicly available data and are protected by fair use, emphasizing that they empower innovation and deliver significant benefits to society.
• The Publisher Stance: The sentiment among publishers is increasingly unified: there is no such thing as a "free" dataset. Whether through class-action lawsuits, individual litigation, or the formation of industry-wide alliances like the News/Media Alliance, publishers are signaling that they will no longer allow their intellectual property to be treated as "free raw material" for multi-billion dollar tech companies.

As of late 2026, the situation remains in flux. While some organizations have settled, such as the Brazilian newspaper Folha, which transitioned from a lawsuit against OpenAI to a commercial agreement, others are digging in for a protracted legal battle.

A System in Transition

The tension between AI development and media sustainability is not merely a legal dispute; it is a fundamental challenge to the future of the information economy. If journalism is the "central pillar of democracy," as many publishers argue, then the current mechanism of AI training risks eroding that pillar for the sake of technological expediency.

For the time being, the industry is split. Some publishers are betting on a "cooperation model," hoping to secure a seat at the table and a share of the AI-generated revenue. Others are betting on the courts to enforce a strict protection of intellectual property, effectively forcing AI companies to pay for the content they consume.

One thing is certain: the era of "tacky tech trafficking," as News Corp CEO Robert Thomson famously described it, is under intense scrutiny. As legal precedents are set and commercial agreements become more common, the relationship between AI platforms and news publishers will define the quality and accessibility of information for the next generation. Whether that future involves a symbiotic partnership or a landscape of litigation remains the central, unresolved question of the digital age.