Meta vorbește cu marile editori de știri pentru a-și licența conținutul pentru instrumentele sale AI. Reports on September 18 confirmed Meta has approached firms like Axel Springer, Fox Corp, and News Corp. The goal is to legally use news articles to Instruiți-i modelele AI.
Aceasta face parte dintr-o tendință mai largă a industriei. Firmele AI plătesc acum date pentru a evita procesele costisitoare ale drepturilor de autor. Mișcarea arată, de asemenea, o modificare a strategiei Meta.
Acum cumpără mai multă tehnologie exterioară după ce propria divizie AI s-a confruntat cu întârzieri. These talks show how tech giants are racing to secure the data that fuels their AI ambitions.
A Strategic Shift From Scraping to Licențiere
Discuțiile semnalează o schimbare majoră în abordarea industriei tehnologice a datelor de formare AI. Ani de zile, modelul predominant a fost de a zgâria cantități mari de date publice fără permisiunea explicită. Now, facing a barrage of litigation, companies are pivoting to licensing as a risk mitigation strategy.
OpenAI has already inked deals with publishers like The Atlantic and Vox Media. Even The New York Times, which is actively suing OpenAI and Microsoft for copyright infringement, struck a licensing agreement with Amazon for its AI products in May.
This dual strategy of litigation and licensing shows the complex calculations media companies are making.
For publishers, these deals represent a critical new revenue stream. Pe măsură ce modelele tradiționale de afaceri se confruntă cu presiunea, conținutul de licențiere pentru dezvoltatorii AI oferă o modalitate de a monetiza arhivele și de a finanța jurnalismul viitor. It is a pragmatic response to an unavoidable technological shift.
The situation is complex, with some firms like Perplexity AI facing accusations of misuse while also attempting to build bridges through revenue-sharing. This highlights the industry’s struggle to find a sustainable and ethical model that satisfies both tech innovators and content creators.
The industry’s legal troubles are mounting. Disney and Universal filed a sweeping lawsuit against AI image generator Midjourney in June 2025, accusing it of building its models on stolen intellectual property.
Disney’s general counsel, Horacio Gutierrez, stated bluntly, “piracy is piracy. And the fact that it’s done by an A.I. company does not make it any less infringing.”
Meta’s ‘Buy Over Build’ Pivot Amid Internal AI Struggles
For Meta, the negotiations also highlight a significant internal pivot. The company is increasingly opting to “buy”rather than “build”its AI capabilities, a direct response to internal development hurdles and intense market pressure.
This strategic shift follows a period of turmoil within Meta’s AI division. Compania și-a revizuit recent laboratoarele AI pentru a doua oară în luni și și-a amânat modelul emblematic Llama 4.
In a recent post, Meta’s Chief AI Officer Alexandr Wang explained the new direction, stating that success requires “to ensure Meta is able to deliver the best possible products for people it will require taking an all-of-the-above approach.”
This philosophy is evident in its recent deals. În august, Meta a licențiat tehnologia puternică a imaginii și video a Midjourney. Just weeks later, it signed a reported $140 million deal with German AI startup Black Forest Labs.
Navigating the High-Stakes Copyright Minefield
The entire generative AI sector is being reshaped by a series of high-stakes court battles. These cases are forcing a legal reckoning over the “fair use”doctrine, which AI firms have long relied on to defend their data scraping practices.
A landmark ruling in a case against AI firm Anthropic has drawn a critical distinction. A judge described the technology as transformative but ruled that acquiring training data from pirated sources was an ‘original sin’ that amounted to theft. Aceasta separă legalitatea procesului AI de actul inițial al achiziției de date.
Această nuanță legală este crucială. As Judge Vince Chhabria remarked in a separate case, “the question, as the courts tell us over and over again, is not whether something is messed up but whether it’s copyright infringement.”
This perspective is gaining traction and puts the onus on AI companies to prove their data was sourced legally, not just that their final product is innovative.
The evolving legal and commercial environment is forcing all players to adaptați-vă. Content platforms like Reddit are now pushing to replace their flat-fee data deals with Google and OpenAI.
They are now negotiating for ‘dynamic pricing’ models that better reflect the ongoing value their user-generated content provides to AI systems.
Meanwhile, publishers are asserting the value of their journalism. As New York Times CEO Meredith Kopit Levien said of her company’s deal with Amazon, “the deal is consistent with our long-held principle that high-quality journalism is worth paying for.”Acest sentiment răsună acum prin săli de consiliu, deoarece companiile media și dezvoltatorii AI negociază viitorul conținutului.
