In an era where the relationship between traditional journalism and the emerging Artificial Intelligence (AI) industry is at a breaking point, The New York Times (NYT) has issued one of its harshest public statements to date. The leadership of the iconic institution did not mince words: companies developing Large Language Models (LLMs) are now openly accused of "shameless theft" of intellectual property. This confrontation is not merely about a newspaper's lost profits, but the very future of information in the digital age.
The Ethical and Legal Dimension of "Looting"
According to recent statements from NYT leadership, the practice of "scraping" millions of articles to train models like OpenAI’s ChatGPT or Google’s Gemini constitutes a blatant violation of competitive norms. The argument is simple yet devastating: tech companies are building multi-billion dollar products using the labor, research, and risk of thousands of journalists as free raw material.
"This is not just a technological evolution, but a systematic effort to appropriate the value produced by others," it was noted. The criticism focuses on the fact that AI bots do not merely link to sources; they often reproduce the content verbatim or provide summaries so detailed that they make visiting the publisher’s website unnecessary. This phenomenon, known as "zero-click search," threatens to drain the advertising and subscription revenue streams that sustain quality journalism.
The "Fair Use" Doctrine Under Scrutiny
For their part, AI companies are entrenching themselves behind the legal principle of "Fair Use." They claim that training their models is analogous to a human reading a newspaper to gain knowledge and then creating something new. However, the NYT and other major publishing houses counter this claim, emphasizing that the scale and automated nature of this process far exceed the boundaries of human learning.
- AI training requires the full storage and processing of protected archives.
- The generated outputs often compete directly with the original content.
- There is no opt-out mechanism that has been respected retroactively for data already utilized.
The ongoing legal battle in the US is expected to set a global judicial precedent. If courts rule in favor of publishers, AI companies may be forced to pay billions of dollars in royalties retroactively—a move that could radically alter Silicon Valley's business model.
Democracy and the Cost of Truth
Beyond the financial figures, NYT leadership warns of an existential threat to democracy. Producing reliable news from war zones, investigative journalism that holds power accountable, and analyzing complex social issues require immense resources. If AI companies destroy the publishers' economic model, who will produce the "fuel" for future AI models?
"If the primary source dries up, Artificial Intelligence will end up recycling itself, leading to a digital desert of misinformation and low-quality content," industry analysts emphasize.
We are already observing the phenomenon of "model collapse," where AIs trained on data generated by other AIs begin to exhibit serious errors and hallucinations. Maintaining a healthy media ecosystem is, therefore, essential even for the survival of the technology itself.
Toward a New Social Contract?
Despite the harsh rhetoric, there are signs that a solution may be found through major commercial agreements. Organizations like News Corp, Axel Springer, and the Associated Press have already signed multi-year licensing deals with OpenAI. However, The New York Times seems to be choosing the path of direct confrontation, seeking not just a financial settlement but the establishment of clear rules that will protect intellectual property in perpetuity.
The stake is whether Artificial Intelligence will function as a parasitic organism on the body of journalism or if it can evolve into a partner that respects the value of human creation. The NYT's decision to label the situation as "shameless theft" indicates that the time for polite consultations is over.
