Latest / Neural Newscast / Brave Software AI Scraping Lawsuit [Operational Drift]
Transcript
- [00:06] Margaret Ellis: On November 2, 2026, a coalition of legacy publishers, led by the New York Times and [00:14] Margaret Ellis: the New York Daily News, filed a formal complaint in federal court against Brave Software. [00:22] Margaret Ellis: The filing alleges the unauthorized and systematic scraping of copyrighted journalism [00:28] Margaret Ellis: on an industrial scale. The legal action highlights a foundational transition in browser behavior. [00:35] Margaret Ellis: A tool once marketed exclusively for consumer privacy now allegedly functions [00:41] Margaret Ellis: as a commercial backdoor for artificial intelligence companies, providing a pathway [00:46] Margaret Ellis: for machines to ingest paywalled content that remains restricted for human readers. [00:51] Margaret Ellis: This investigation examines how complex technical systems quietly drift away [00:56] Margaret Ellis: from their stated intent, oversight, and control, and explores the consequences [01:02] Margaret Ellis: when no single entity is held responsible for the resulting extraction. I'm Margaret Ellis. This is [01:08] Margaret Ellis: Operational Drift. The complaint filed earlier this week represents more [01:15] Margaret Ellis: than a standard copyright dispute between media entities and a tech firm. It is a record of [01:20] Margaret Ellis: how a browser's internal logic has fundamentally shifted. For years, Brave Software [01:26] Margaret Ellis: marketed itself to users as a shield. It promised to block the trackers [01:30] Margaret Ellis: and advertisements that publishers rely on for their primary revenue, framing this intervention [01:35] Margaret Ellis: as a victory for consumer privacy. But the lawsuit filed by the New York Times, the Chicago Tribune, [01:42] Margaret Ellis: and the Orlando Sentinel suggests that the shield has become a sieve designed to filter out [01:47] Margaret Ellis: the publishers' income while siphoning their data. According to the court documentation, [01:53] Margaret Ellis: Brave has not simply been blocking content for the benefit of its users; [01:57] Margaret Ellis: it has been indexing that same content for the benefit of machines. The publishers allege [02:02] Margaret Ellis: that Brave has scraped their websites without permission to build what the company calls [02:06] Margaret Ellis: the Brave Index. This is described as a massive real-time repository of online content. [02:13] Margaret Ellis: The drift here is subtle but foundational: a tool that was once defined by what [02:18] Margaret Ellis: it prevented—data extraction from the agent—is now being defined by what [02:22] Margaret Ellis: it facilitates—the wholesale extraction of intellectual property from the publisher. [02:28] Margaret Ellis: The lawsuit names several of the most prominent players [02:31] Margaret Ellis: in the generative artificial intelligence space as the primary beneficiaries of this index. [02:36] Margaret Ellis: These entities include Anthropic, Perplexity, Mistral, XAI, and Cohere. [02:42] Margaret Ellis: The central allegation is that these companies pay Brave for access to this curated index, [02:47] Margaret Ellis: using it to train their massive models and to provide real-time responses to user queries using [02:51] Margaret Ellis: current news data. In this arrangement, Brave acts as a broker, selling the work of journalists [02:57] Margaret Ellis: to the very companies that are positioned to replace the traditional model of news consumption [03:01] Margaret Ellis: and distribution. Graham James, a spokesperson for the New York Times, [03:06] Margaret Ellis: stated that Brave has flouted established copyright law and ignored repeated requests to stop [03:13] Margaret Ellis: its infringing conduct. The core of the Times' argument is [03:17] Margaret Ellis: that Brave is essentially stealing content to power its own proprietary products, [03:22] Margaret Ellis: and then providing that same stolen content to the world's most well-resourced [03:27] Margaret Ellis: artificial intelligence companies. This is not a failure of the system's code; [03:32] Margaret Ellis: it is an intentional pivot in how that code is monetized. The infrastructure designed to navigate [03:38] Margaret Ellis: the web has been repurposed to harvest it. One of the most specific points in the complaint involves [03:45] Margaret Ellis: the paywall blocker feature built into the Brave browser. The software promotes [03:50] Margaret Ellis: that it can provide access to hundreds of top publishers by bypassing their subscription gates. [03:55] Margaret Ellis: To a casual user, this looks like a useful utility. To a publisher, it looks like a direct assault [04:01] Margaret Ellis: on their business model. But for the AI companies, it is something else entirely. [04:06] Margaret Ellis: It is a way to ingest premium, high-quality data that is otherwise protected by a financial barrier. [04:13] Margaret Ellis: The Drift has converted a browser feature meant for the agent into a corporate exploit [04:18] Margaret Ellis: for training sets. Stephen Lieberman, the attorney who filed the lawsuit on behalf of [04:23] Margaret Ellis: the Chicago Tribune and other news organizations, uses a specific term to describe this behavior. [04:29] Margaret Ellis: He calls it laundering pirated reporting. He argues that buying stolen goods secondhand [04:35] Margaret Ellis: does not make them legitimate in the eyes of the law. This framing suggests that Brave is serving as [04:40] Margaret Ellis: a technical middleman that allows artificial intelligence giants to avoid being caught red-handed [04:46] Margaret Ellis: while they access the journalism they have no legal right to use. It is a calculated relocation [04:51] Margaret Ellis: of liability through a technical intermediary. When journalists at The Verge reached out [04:56] Margaret Ellis: for comment, the responses from the technology sector were telling. While many [05:01] Margaret Ellis: of the artificial intelligence companies remained silent, [05:04] Margaret Ellis: Perplexity's chief communications officer Jesse Dwyer dismissed the suit [05:08] Margaret Ellis: as containing disinformation. This reaction points to a deepening divide between those who produce [05:13] Margaret Ellis: the record and those who process it. There is no disagreement that the data is being used. [05:19] Margaret Ellis: The disagreement is over whether the way it was acquired matters once it has been integrated into [05:23] Margaret Ellis: a model's weights or a search engine's index. In the context of operational drift, [05:29] Margaret Ellis: we have to look at the incentives that allowed this to happen. The artificial intelligence industry [05:35] Margaret Ellis: has an insatiable need for fresh, accurate, and human-verified data. As publishers have tightened [05:43] Margaret Ellis: their paywalls to survive, the supply of free, high-quality data on the open web has dwindled. [05:50] Margaret Ellis: This created a significant market gap. Brave software, possessing an index of the web [05:57] Margaret Ellis: and a browser that could bypass those paywalls, sat directly in that gap. [06:02] Margaret Ellis: The drift toward becoming a data provider for artificial intelligence was a predictable response [06:08] Margaret Ellis: to economic pressure. The publishers note that Brave has never asked for permission to access [06:14] Margaret Ellis: their websites for this specific purpose, nor have they offered any form of compensation. [06:20] Margaret Ellis: The complaint states that Brave has taken active steps to prevent publishers from blocking [06:25] Margaret Ellis: the browser's access to their servers. This creates a technical stalemate. [06:30] Margaret Ellis: If a publisher tries to protect its content, the browser's code is updated to circumvent [06:35] Margaret Ellis: that protection. What began as a tool for user agency has evolved into a tool that overrides [06:41] Margaret Ellis: the agency of the content creator and the copyright owner. We should consider the timeline [06:47] Margaret Ellis: of Brave's development. It entered the market as a principled alternative [06:51] Margaret Ellis: to the surveillance capitalism of major tech firms like Google. Its pivot to selling the Brave Index [06:57] Margaret Ellis: suggests that even the alternatives are susceptible to the gravity [07:01] Margaret Ellis: of artificial intelligence training needs. When a company owns a massive index of the internet, [07:07] Margaret Ellis: the temptation to monetize that index as a training set is immense. It is a shift from serving [07:12] Margaret Ellis: the person behind the keyboard to serving the algorithm behind the query, [07:17] Margaret Ellis: often without the agent knowing they are participating. The Brave Index is described in [07:22] Margaret Ellis: the legal filing as being monetized by providing "real-time online content." For companies building [07:29] Margaret Ellis: artificial intelligence, real-time access is the gold standard. It allows a model to discuss events [07:35] Margaret Ellis: that happened 10 minutes ago, rather than events from two years ago. By providing this, [07:40] Margaret Ellis: Brave isn't just selling a static database; it is selling a live feed of global journalism. [07:47] Margaret Ellis: This live feed is the product of thousands of journalists, editors, and photographers [07:51] Margaret Ellis: whose work is being bundled and resold as a commodity without their consent or knowledge. [07:57] Announcer: This case brings into sharp focus the concept of regulatory compression. [08:03] Announcer: As artificial intelligence technology moves faster than the slow machinery of the law, [08:08] Announcer: companies find gray zones where they can operate with impunity. Brave's defense will likely hinge on [08:15] Announcer: how it characterizes [08:16] Announcer: its index and its paywall blocker feature. If it can convince the court it is merely [08:22] Margaret Ellis: a neutral tool for [08:23] Announcer: users, it may avoid liability. [08:27] Margaret Ellis: But the publishers [08:27] Announcer: are arguing [08:28] Margaret Ellis: that Brave is not neutral. They are arguing that Brave is an active, profitable participant [08:34] Margaret Ellis: in a system of extraction that it has intentionally built and marketed. [08:40] Margaret Ellis: Let us look closer at the companies mentioned as clients of this index. Anthropic [08:45] Margaret Ellis: and Perplexity are often cited as leaders in ethical or responsible artificial intelligence. Yet, [08:52] Margaret Ellis: if the allegations in this suit are accurate, they are relying on a secondary market for data [08:57] Margaret Ellis: that sidesteps the ethical requirements of direct licensing. This is how drift becomes systemic [09:03] Margaret Ellis: across an entire industry. One company provides the backdoor, and other companies use it, [09:09] Margaret Ellis: creating a chain of custody that makes it increasingly difficult to pinpoint exactly [09:14] Margaret Ellis: where the legal line was crossed. In March of 2024, the conversation around artificial intelligence [09:22] Margaret Ellis: and copyright was focused on whether training was fair use. By November of 2026, [09:28] Margaret Ellis: the conversation has moved to the infrastructure of the scraping itself. [09:33] Margaret Ellis: The brave lawsuit suggests that the industry has moved past the question of [09:37] Margaret Ellis: whether they should use this data, and into a phase where they are finding [09:40] Margaret Ellis: increasingly sophisticated ways to acquire it quietly. The method of acquisition has become [09:46] Margaret Ellis: the primary point of contention in the federal courts. According to the complaint, [09:52] Margaret Ellis: Brave also uses this index to respond to specific user queries within its own search products. [09:58] Margaret Ellis: This creates a recursive loop of information. The browser scrapes the news, [10:03] Margaret Ellis: uses that news to build an index, sells that index to AI companies, and then uses the same index [10:09] Margaret Ellis: to provide generated answers that prevent the agent from ever needing to visit [10:13] Margaret Ellis: the original news site. Each step in this process erodes the link between the person who did [10:19] Margaret Ellis: the work and the person who consumes the information, effectively cutting the publisher out [10:23] Margaret Ellis: of the loop. The publisher's attorney, Stephen Lieberman, emphasized [10:28] Margaret Ellis: that the artificial intelligence giants are desperate to avoid being caught stealing. [10:34] Margaret Ellis: This desperation acts as a powerful market incentive. It creates a space for companies [10:40] Margaret Ellis: like Brave to act as a buffer. If an AI company can point to a third-party index as its source, [10:47] Margaret Ellis: it adds a layer of separation between its model and the act of scraping. [10:52] Margaret Ellis: This separation is the goal of the arrangement. It is a way to relocate the legal [10:57] Margaret Ellis: and ethical liability of data acquisition to a third party, creating plausible deniability. [11:04] Margaret Ellis: We see here a failure of interpretability, not in the model's code, but in the institutional system [11:10] Margaret Ellis: itself. When a browser company says it is protecting privacy, we interpret that to mean [11:16] Margaret Ellis: it is protecting the agent from observation. We do not necessarily interpret it to mean [11:22] Margaret Ellis: it is building a proprietary database of the agent's destination. The drift happens in the gap [11:28] Margaret Ellis: between our interpretation of a service and the technical reality of what that service is collecting [11:33] Margaret Ellis: and selling behind the scenes. Brave has not yet issued a full formal response [11:39] Margaret Ellis: to the specific allegations of laundering. But the tension is clear to anyone following the case. [11:45] Margaret Ellis: The company has moved from being a niche privacy tool to a central node [11:49] Margaret Ellis: in the artificial intelligence supply chain. This is the nature of operational drift. [11:55] Margaret Ellis: It rarely happens through a single loud announcement. It happens through the gradual accumulation [12:00] Margaret Ellis: of new features, an index here, a paywall bypass there, until the original intent of the system is [12:07] Margaret Ellis: unrecognizable to the people who first supported it. The lawsuit also mentions that Brave has taken [12:13] Margaret Ellis: technical steps that prevent publishers from blocking the browser's access to their websites. [12:18] Margaret Ellis: This is a crucial detail for the court to consider. It indicates that this is not [12:22] Margaret Ellis: a passive observation of the public web, but an active, persistent effort to maintain [12:27] Margaret Ellis: a data pipeline even against the express wishes of the content owners. [12:32] Margaret Ellis: It is the technical enforcement of a decision [12:34] Margaret Ellis: that no one outside of Brave Software's board of directors ever agreed to or authorized. [12:40] Announcer: If we trace the continuity of this behavior, we see a clear path leading from ad blocking [12:47] Announcer: to revenue blocking to data harvesting. Each step along this path was framed as a benefit [12:54] Announcer: for the end user. Ad blocking makes the site load faster. Revenue blocking through paywall bypassing [13:01] Announcer: makes the site cheaper to access. Data harvesting makes the artificial intelligence smarter. [13:09] Margaret Ellis: But at the end of this path, the publishers who create the original value in the first place [13:15] Margaret Ellis: are left with no way to sustain their operations or pay their reporters. The legal filing by [13:21] Margaret Ellis: the New York Daily News, the Chicago Tribune, and the Orlando Sentinel [13:26] Margaret Ellis: marks a moment where the producers of information are attempting to reassert control [13:31] Margaret Ellis: over the boundaries of their digital property. They are identifying Brave as the point of failure [13:37] Margaret Ellis: in the copyright chain. They are arguing that the browser, which was once a tool for navigating [13:43] Margaret Ellis: the internet, has become a tool of industrial-scale infringement designed to benefit a new class [13:49] Margaret Ellis: of technology conglomerates. In previous reporting, the focus has often been on the black box [13:56] Margaret Ellis: of the artificial intelligence model itself. This lawsuit shifts the focus to the black box [14:02] Margaret Ellis: of the data supply chain. We are beginning to see how the sausage is made, [14:07] Margaret Ellis: and it involves a complex network of indices, scrapers, and brokers that operate in the shadows [14:13] Margaret Ellis: of the open web. Brave software is now the primary subject of this investigation [14:19] Margaret Ellis: into that supply chain, as publishers demand transparency regarding how their archives [14:23] Margaret Ellis: are being used. The implication for the future of the web is significant. [14:29] Margaret Ellis: If browsers can successfully claim that their indexing of content is a protected activity, [14:35] Margaret Ellis: then the paywall—the primary survival mechanism for modern journalism—becomes effectively obsolete [14:41] Margaret Ellis: for machines, even as it remains in place for human readers. This creates a two-tiered web [14:48] Margaret Ellis: where artificial intelligence systems have more access than the people they are designed to serve, [14:53] Margaret Ellis: further centralizing power in the hands of those who control the indices. [14:57] Announcer: Operational drift often manifests as the normalization of a previously unacceptable behavior. [15:04] Announcer: In 2018, the idea of a browser selling its index of paywalled sites to third-party companies [15:12] Announcer: would have been seen as a massive breach of trust. By 2026, [15:17] Announcer: it is presented by some in the tech industry as a standard data service. The disinformation claim [15:23] Margaret Ellis: by Perplexity spokesperson suggests that they view this as a legitimate, perhaps even necessary, [15:30] Margaret Ellis: part of the artificial intelligence ecosystem that is now being unfairly challenged. We must ask: [15:37] Margaret Ellis: who noticed this change as it was happening? The publishers noticed when their traffic patterns [15:42] Margaret Ellis: and subscription numbers no longer aligned with the value they were providing. [15:47] Margaret Ellis: The AI companies noticed when they found a convenient paid source for real-time news data. [15:53] Margaret Ellis: But for the average user of the Brave browser, the drift was likely invisible. [15:58] Margaret Ellis: They were still just using a browser that blocked ads. They were entirely unaware that their browser [16:03] Margaret Ellis: was also a silent partner in a global data extraction project of this magnitude. [16:09] Margaret Ellis: The complaint filed on November 2nd details how Brave has never compensated the publishers [16:14] Margaret Ellis: for this use of their copyrighted content. This lack of compensation is not a simple oversight; [16:21] Margaret Ellis: it is a critical feature of the business model. To pay for the content would be to acknowledge [16:25] Margaret Ellis: that the content has an inherent value that belongs to someone else. By not paying, [16:31] Margaret Ellis: Brave maintains the convenient fiction that it is simply providing a technical service rather [16:35] Margaret Ellis: than selling a product derived from the work of others. The role of Anthropic, Mistral, [16:41] Margaret Ellis: and others in this lawsuit is also critical to understand. These companies have raised [16:46] Margaret Ellis: billions of dollars on the promise of transforming human information. If it is proven in court [16:51] Margaret Ellis: that they are knowingly purchasing stolen data through Brave, it challenges the popular narrative [16:56] Margaret Ellis: of artificial intelligence as an emergent clean technology. It suggests that the intelligence [17:02] Margaret Ellis: in these systems [17:03] Margaret Ellis: is in part a product of institutionalized piracy hidden behind a sophisticated technical interface. [17:10] Announcer: The drift here has relocated accountability into a gray area between the browser, the index, [17:17] Announcer: and the model. If an agent receives a copyrighted excerpt from an AI, is the AI company responsible? [17:25] Margaret Ellis: Is the index [17:26] Announcer: provider Brave responsible? [17:29] Margaret Ellis: Or is the entire system designed specifically so that no one person [17:33] Margaret Ellis: or entity can be held fully accountable for the final outcome? This diffusion of responsibility is [17:41] Margaret Ellis: the core of the problem, allowing each actor to point the finger at another while the extraction [17:47] Margaret Ellis: continues. As of today, November 7th, 2026, the lawsuit is still in its earliest stages. [17:57] Margaret Ellis: We are looking at a record that is still being written by the court, but the signal is clear: [18:03] Margaret Ellis: the boundary between a browser and a data broker has dissolved. What we are witnessing is [18:09] Margaret Ellis: the quiet transformation of the infrastructure of the internet into a dedicated feeder system [18:14] Margaret Ellis: for artificial intelligence, with the browser serving as the primary extraction point [18:19] Margaret Ellis: for information. We have moved from a web of links to a web of indices. In a web of links, [18:27] Margaret Ellis: the agent travels to the source to find what they need. In a web of indices, [18:33] Margaret Ellis: the source is harvested, processed, and resold before the agent ever sees it. Brave Software, [18:40] Margaret Ellis: through the Brave Index, has positioned itself as the processor of that world. [18:46] Margaret Ellis: The publisher's lawsuit is a desperate attempt to force the system back into a model [18:51] Margaret Ellis: where the source of the information still matters and possesses legal rights. [18:56] Margaret Ellis: This drift was already happening well before the filing. Today is just when the New York Times [19:01] Margaret Ellis: and its peers decided to officially notice it in a court of law. The documentation provided in [19:08] Margaret Ellis: the complaint serves as a map of the territory we have entered—a territory where privacy features [19:13] Margaret Ellis: are used as cover for data collection, and where laundering is a viable business strategy [19:18] Margaret Ellis: for the world's most advanced technology firms seeking to dominate the future of information. [19:25] Margaret Ellis: The attorney Stephen Lieberman noted that buying stolen goods secondhand doesn't make [19:29] Margaret Ellis: them legitimate. This is a simple moral and legal principle that is currently being tested by [19:35] Margaret Ellis: incredibly complex technical systems. If the law accepts Brave's model, [19:40] Margaret Ellis: it essentially signals that any intermediary can scrape and sell any data, [19:44] Margaret Ellis: so long as they call it an index and provide it to an AI. It would represent [19:49] Margaret Ellis: a final collapse of the traditional understanding of intellectual property in the digital age. [19:54] Announcer: We will continue to monitor the filings and the technical shifts in this case. [20:00] Announcer: The transition of Brave from a consumer privacy tool to an AI infrastructure middleman is [20:06] Announcer: a textbook example of how the stated intent of a system can be overruled by the incentives [20:12] Announcer: of the market it inhabits. The drift is not an accident. It is the result of a series of logical, [20:18] Announcer: profitable choices made in the absence of oversight, where the value of data outweighs the values [20:25] Announcer: of the creator. The question that remains is whether the legal system can keep up [20:30] Announcer: with the speed at which these technical definitions are being rewritten by the tech industry. [20:35] Margaret Ellis: If an index is no longer a guide to information but a total replacement for it, [20:40] Margaret Ellis: then our existing copyright frameworks are essentially obsolete. We do not yet know [20:45] Margaret Ellis: who will be held responsible for the billions of words already ingested through this backdoor, [20:51] Margaret Ellis: or if the damage to the news industry is already irreversible. Does the purchase of a curated index [20:57] Margaret Ellis: from a third-party broker legally absolve an AI company of the responsibility to know [21:03] Margaret Ellis: where its training data came from? Or is the browser simply the latest mechanism used to hide [21:08] Margaret Ellis: the fact that the emergent behavior of AI still relies on the uncompensated labor [21:13] Margaret Ellis: of human reporters? For more on this case, visit operationaldrift.neuralnewscast.com. [21:21] Margaret Ellis: Neural Newscast is AI-assisted, human-reviewed. [21:25] Margaret Ellis: View our AI transparency policy at neuralnewscast.com. This record is closed.