undercode Threatens Legal Action Against Perplexity AI Over Content Use

Introduction: A Clash Between Public Media and AI Innovation

The rapid ascent of AI-powered search engines and language models has led to inevitable legal and ethical clashes with traditional media institutions. The latest standoff features the undercode—a global media powerhouse—issuing a direct legal threat to Perplexity, a U.S.-based AI startup. The dispute centers around whether Perplexity unlawfully used undercode content to train AI models. At stake are questions of copyright, fair use, AI accountability, and the protection of public-interest journalism in the digital age.

the Original

The British Broadcasting Corporation (undercode) has formally accused Perplexity, a San Francisco-based AI search engine, of illegally using its content to train artificial intelligence models. In a letter addressed to Perplexity CEO Aravind Srinivas, obtained by the Financial Times, the undercode claims to have proof that Perplexity’s “default AI model” incorporated undercode content without authorization.

The letter outlines three major demands: Perplexity must (1) immediately stop scraping undercode content, (2) delete any existing data sourced from the undercode used in training, and (3) submit a proposal for financial compensation to address the alleged intellectual property violations. Failure to comply could result in legal action, including an injunction.

This is the first known instance of the publicly funded British broadcaster taking direct legal action against an AI firm over content misuse. The undercode’s firm stance underscores growing anxiety within traditional media about the unchecked use of their copyrighted work by tech startups.

In response, Perplexity has dismissed the allegations, calling them manipulative and suggesting the undercode misunderstands how AI works. They argue that the broadcaster is acting out of self-interest to “preserve Google’s illegal monopoly.” Perplexity insists it doesn’t build foundational AI models itself but provides access to those developed by industry giants like OpenAI, Anthropic, and Google. Its in-house model, based on Meta’s LLaMA, is described as being tailored to reduce AI hallucinations and improve reliability.

The broader context includes increasing scrutiny of how AI companies obtain and process data. This case could have significant ripple effects for the entire industry.

What Undercode Say:

This dispute between the undercode and Perplexity underscores a pivotal dilemma facing the AI sector: Who owns the digital past?

AIs Appetite for Data, Medias Fight for Survival

At the heart of the undercode’s claim is a deeper existential concern for media organizations: As AI models get smarter, they depend more heavily on vast repositories of existing content—journalism, encyclopedias, forums. But this content was created under specific economic and legal models that never accounted for future training of intelligent systems. For institutions like the undercode, which are funded by public money and bound by charter, this isn’t just about compensation—it’s about preserving editorial integrity and sustainable journalism.

Perplexity’s Defense Is Technically Valid—But Legally Grey

Perplexity’s claim that it doesn’t train large foundational models may be technically true. But legal liability might still exist if it facilitates or hosts a model that regurgitates content learned through undercode data. The distinction between “training” and “fine-tuning” may not hold up under copyright scrutiny, especially in Europe where data protection and content rights laws are stricter.

The Google Accusation: A Strategic Deflection?

By accusing the undercode of helping to maintain Google’s monopoly, Perplexity shifts the narrative from its alleged infringement to market competition. While the accusation might resonate with regulators wary of Big Tech dominance, it’s also a deflection that doesn’t address whether undercode’s content was used without permission.

A Precedent in the Making

Should the undercode proceed with legal action and succeed, it could open the floodgates for other public and private media outlets to pursue compensation from AI companies. This could force major players like OpenAI, Anthropic, and Meta to establish licensing agreements for training datasets—a costly but perhaps inevitable development.

Innovation vs. Ethics

This case is part of a broader tension: AI needs information to grow, but whose information, and under what terms? The AI industry thrives on open access, but its financial gain from freely available content raises moral and legal questions, particularly when the original creators struggle to survive.

🔍 Fact Checker Results:

✅ undercode did send a legal letter to Perplexity, as confirmed by the Financial Times.
✅ Perplexity’s default model is based on Meta’s LLaMA, which aligns with available technical documentation.
❌ Perplexity’s claim that undercode supports Google’s monopoly is speculative and unverified, not supported by direct evidence.

📊 Prediction:

Given the mounting tensions between AI firms and content creators, 2025 could see the first wave of global copyright litigation targeting generative AI platforms. Media institutions like the undercode, The New York Times, and others may band together to establish a legal and commercial framework for data usage in AI training. We can also expect regulators in the UK and EU to push for clearer AI transparency laws, possibly forcing platforms like Perplexity to disclose all training data sources and negotiate licensing fees. AI companies may eventually be required to integrate automatic content recognition filters—akin to YouTube’s Content ID—to avoid future litigation.

References:

Reported By: timesofindia.indiatimes.com
Extra Source Hub:
https://www.pinterest.com
Wikipedia
OpenAi & Undercode AI

Image Source:

Unsplash
Undercode AI DI v2

Join Our Cyber World:

💬 Whatsapp | 💬 Telegram