Meta Delays Controversial AI-Training Business Terms Indefinitely Amid Publisher Backlash
Meta pushes back implementation of updated terms that could permit AI model training on publisher content, citing need for extended industry consultation.
Last verified:
Meta Extends Delay on AI-Training Business Terms Indefinitely
Meta will delay implementation of its updated business terms beyond the November 20 deadline originally set, according to Meta communications head Andy Stone, who announced on October 28 that the company intends to conduct extended consultation with industry stakeholders. According to The Verge, the indefinite postponement reflects growing backlash from publishers concerned that the revised terms could still permit Meta to train its AI models on news content hosted on Facebook and Instagram.
The Evolution of Meta’s Terms and Legal Ambiguity
The controversy began with business terms scheduled for October 30 that appeared to grant Meta explicit permission to “train” its AI products on content from publisher websites using its platforms. According to The Verge, Meta later removed that training language, but Nieman Lab’s legal experts determined the updated version does not fully eliminate the possibility of AI training. This residual ambiguity—whether Meta could train models on publisher content under different contractual language—has become the focal point of publisher concern and regulatory scrutiny.
Bipartisan and Industry Opposition
The pushback extends beyond publishers. According to The Verge, U.S. Senator Marsha Blackburn (R-TN) characterized the updated terms as Meta “attempting to exploit American creators and businesses,” while U.S. Senator Adam Schiff (D-CA) called for Meta to reverse the change entirely. The News/Media Alliance has also mounted formal opposition, contributing to Stone’s statement that Meta is “committed to getting this right” and will delay implementation pending full industry dialogue.
Why This Matters
The indefinite delay signals that Meta faces genuine friction on AI training consent from both organized publishers and elected officials across the political spectrum. For news publishers deciding whether to maintain their presence on Meta’s platforms, the extended timeline offers temporary protection but leaves the core question—whether Meta can train on their content—unresolved. If Meta ultimately revises the terms to explicitly prohibit AI training without separate licensing agreements, it could set precedent for how other platforms negotiate AI-content access. Conversely, if Meta proceeds with language that preserves training rights, publishers may face pressure to negotiate separate opt-out or licensing arrangements, reshaping the economic relationship between platforms and news organizations.
Frequently Asked Questions
What are Meta's new business terms about?
The updated terms govern how Meta can use content from news publishers on Facebook and Instagram, specifically whether Meta can train its AI models on that content. Meta removed explicit 'training' language after backlash, but legal experts say ambiguity remains.
Who is opposing these terms?
Publishers represented by the News/Media Alliance, bipartisan U.S. senators including Marsha Blackburn (R-TN) and Adam Schiff (D-CA), and legal experts have raised concerns about potential AI training on publisher content without explicit consent.
When will Meta implement the new terms?
Implementation is now indefinitely delayed pending industry consultation. The original October 30 deadline was moved to November 20; Meta now says the consultation may extend beyond that date.