[{"data":1,"prerenderedAt":729},["ShallowReactive",2],{"article-/newsroom/project-gutenberg-ai-semantic-book-discovery":3,"related-/newsroom/project-gutenberg-ai-semantic-book-discovery":177},{"id":4,"title":5,"author":6,"body":7,"date":147,"description":148,"experienceName":104,"experienceUrl":106,"extension":149,"faqs":150,"image":169,"lastModified":6,"meta":170,"navigation":171,"path":172,"seo":173,"seoDescription":6,"seoTitle":6,"stem":174,"tags":6,"topic":175,"__hash__":176},"newsroom/newsroom/project-gutenberg-ai-semantic-book-discovery.md","Project Gutenberg AI: Discovering Books by What They Actually Mean",null,{"type":8,"value":9,"toc":135},"minimark",[10,27,30,35,38,41,45,51,56,69,73,76,80,83,86,93,101,107,111,117,120,127],[11,12,13,14,19,20,26],"p",{},"Project Gutenberg AI is Empathy AI's intelligent book discovery system, built on our ",[15,16,18],"a",{"href":17},"/newsroom/introducing-knowledge-engine/","Knowledge Engine"," and developed in collaboration with ",[15,21,25],{"href":22,"rel":23},"https://www.gutenberg.org/",[24],"nofollow","Project Gutenberg",", the world's oldest digital library. It categorizes and recommends literature based on deep semantic analysis of actual book content, not just titles, genres, author names, or publisher metadata.",[11,28,29],{},"Where Project Gutenberg has spent over 50 years making public domain literature freely accessible (75,000+ eBooks and counting), Project Gutenberg AI adds a new layer: the ability to discover those works by what they actually mean. Themes, emotions, narrative structures, philosophical undercurrents. Content discovery that goes beyond keywords, processing what books actually say rather than what labels have been attached to them. And it runs entirely on Empathy AI's private, self-hosted infrastructure.",[31,32,34],"h2",{"id":33},"why-traditional-book-discovery-fails-readers","Why Traditional Book Discovery Fails Readers",[11,36,37],{},"Most book discovery tools rely on metadata: genre tags, author name matching, bestseller lists, and \"customers also bought\" algorithms trained on purchase behavior. According to research published in the Journal of Documentation, metadata-based recommendation systems achieve relevance rates below 40% for readers seeking thematic or emotional connections with their next book.",[11,39,40],{},"A reader searching for \"a quiet story about grief and resilience\" will not find what they need through genre filters. Metadata does not capture what a book feels like to read. Content analysis does.",[31,42,44],{"id":43},"how-project-gutenberg-ai-works","How Project Gutenberg AI Works",[11,46,47,48,50],{},"Project Gutenberg AI is powered by Empathy AI's ",[15,49,18],{"href":17},", an Agentic RAG (Retrieval-Augmented Generation) platform that transforms unstructured content into semantically searchable knowledge. The same contextual retrieval and enrichment pipeline that makes Knowledge Engine effective for enterprise documentation is applied here to literature, analyzing books at the content level through three layers of semantic processing:",[52,53,55],"h3",{"id":54},"deep-content-analysis","Deep Content Analysis",[11,57,58,59,63,64,68],{},"The system ingests the full text of books from the ",[15,60,62],{"href":22,"rel":61},[24],"Project Gutenberg catalogue"," and processes narrative structure, thematic patterns, emotional arcs, character dynamics, and stylistic elements. This goes far deeper than traditional natural language processing keyword extraction. Using Empathy AI's ",[15,65,67],{"href":66},"/newsroom/why-we-only-use-open-source-llms/","open-source LLMs"," running on the Knowledge Engine's contextual retrieval pipeline, the system identifies what a book is about at a semantic level, not just what words it contains.",[52,70,72],{"id":71},"intent-matching","Intent Matching",[11,74,75],{},"When a reader describes what they are looking for, using moods, themes, life moments, or emotional states, Project Gutenberg AI matches that intent against its deep content index. The result is recommendations that feel personally relevant, not algorithmically obvious.",[31,77,79],{"id":78},"content-discovery-not-behavior-tracking","Content Discovery, Not Behavior Tracking",[11,81,82],{},"Most book recommendation engines rely on collaborative filtering: tracking what other readers purchased, browsed, or rated. This approach has two fundamental problems.",[11,84,85],{},"First, it creates filter bubbles. Readers see variations of what they have already consumed, not genuinely new discoveries. Second, it requires surveillance: monitoring reading behavior, purchase history, and browsing patterns to fuel the recommendation engine.",[11,87,88,92],{},[89,90,91],"strong",{},"Project Gutenberg AI needs neither",". Recommendations are based on what books contain, not on what readers do. Your reading behavior is not the product. The books themselves are the signal.",[11,94,95,96,100],{},"All processing runs on Empathy AI's ",[15,97,99],{"href":98},"/newsroom/net-zero-bioclimatic-building/","private GPU infrastructure",". No reader data is shared with external platforms, no behavior is tracked for advertising purposes, and no reading history is used to train third-party models.",[102,103],"experience-cta",{"name":104,"slug":105,"url":106},"Project Gutenberg AI","project-gutenberg-ai-semantic-book-discovery","https://projectgutenberg.empathy.ai",[31,108,110],{"id":109},"from-gutenberg-to-discovery","From Gutenberg to Discovery",[11,112,113,116],{},[15,114,25],{"href":22,"rel":115},[24]," was founded in 1971 by Michael S. Hart, making it the world's oldest digital library. For over 50 years, thousands of volunteers have digitized and proofread public domain literature, building a freely accessible collection of more than 75,000 eBooks. It was the original open-access revolution for books, decades before the internet made it obvious.",[11,118,119],{},"The challenge Project Gutenberg faces today is not availability. The books are there, free and open. The challenge is discovery. With 75,000 works spanning centuries of literature, finding the right book still depends on knowing what you are looking for: a title, an author, a subject heading. Readers with broader or more exploratory intent (\"something that captures the same existential weight as Dostoevsky but in a shorter format\") have no path forward through traditional search.",[11,121,122,123,126],{},"That is where Empathy AI's collaboration with Project Gutenberg begins. By applying the ",[15,124,125],{"href":17},"Knowledge Engine's"," semantic analysis capabilities to the Gutenberg catalogue, we add a discovery layer that the original library was never designed to have. Readers can now explore literature through meaning, not just metadata.",[11,128,129,130,134],{},"This is the same philosophy behind the broader vision of ",[15,131,133],{"href":132},"/newsroom/de-anthropomorphizing-ai/","AI at the service of genuine empathy",": computational tools that enhance human connection with literature rather than replacing the joy of discovery with algorithmic prediction. Project Gutenberg gave the world free access to books. Project Gutenberg AI helps readers find the ones that matter to them.",{"title":136,"searchDepth":137,"depth":137,"links":138},"",2,[139,140,145,146],{"id":33,"depth":137,"text":34},{"id":43,"depth":137,"text":44,"children":141},[142,144],{"id":54,"depth":143,"text":55},3,{"id":71,"depth":143,"text":72},{"id":78,"depth":137,"text":79},{"id":109,"depth":137,"text":110},"2026-03-03","Built on the Knowledge Engine and in collaboration with Project Gutenberg, Project Gutenberg AI brings semantic book discovery to 75,000+ public domain works.","md",[151,154,157,160,163,166],{"question":152,"answer":153},"What is Project Gutenberg AI?","Project Gutenberg AI is Empathy AI's intelligent book discovery system, built on the Knowledge Engine (our Agentic RAG platform) and developed in collaboration with Project Gutenberg. It analyzes the actual content of over 75,000 public domain books to help readers find literature that resonates with their interests, rather than relying on genre tags or purchase behavior.",{"question":155,"answer":156},"How is this different from Amazon or Goodreads recommendations?","Amazon and Goodreads primarily use collaborative filtering based on purchase and rating behavior. Project Gutenberg AI analyzes what books actually contain at a semantic level, enabling discovery based on meaning and emotional connection rather than behavioral tracking.",{"question":158,"answer":159},"Does Project Gutenberg AI track reading behavior?","No. Recommendations are generated from content analysis, not user tracking. All processing happens on Empathy AI's private infrastructure in Asturias, Spain. No reader data is shared with external providers.",{"question":161,"answer":162},"What kinds of queries can Project Gutenberg AI handle?","Readers can describe what they want using natural language: moods, themes, comparisons, or life moments. For example, \"something hopeful but not naive\" or \"books with a similar atmosphere to The Remains of the Day.\"",{"question":164,"answer":165},"What is the relationship with Project Gutenberg?","Project Gutenberg AI is developed in collaboration with Project Gutenberg (gutenberg.org), the pioneering digital library that has been making public domain literature freely accessible since 1971. Empathy AI extends their mission by adding AI-powered semantic discovery to the Gutenberg catalogue, helping readers navigate over 75,000 works through meaning and connection rather than metadata alone.",{"question":167,"answer":168},"Is Project Gutenberg AI available for bookstores and publishers?","Yes. Project Gutenberg AI is designed for organizations in the book and publishing industry that want to offer superior discovery experiences. The same Knowledge Engine technology that powers Project Gutenberg AI can be configured for any literary catalogue. Contact Empathy AI for partnership details.","/media/newsroom/article1_pg.webp",{},true,"/newsroom/project-gutenberg-ai-semantic-book-discovery",{"title":5,"description":148},"newsroom/project-gutenberg-ai-semantic-book-discovery","Product","G_uqIu-gswWtYMRfsQ56MzYA9PUxMQZZ2kVS_LKmA-A",[178,403,571],{"id":179,"title":180,"author":6,"body":181,"date":379,"description":380,"experienceName":18,"experienceUrl":227,"extension":149,"faqs":381,"image":397,"lastModified":6,"meta":398,"navigation":171,"path":399,"seo":400,"seoDescription":6,"seoTitle":6,"stem":401,"tags":6,"topic":175,"__hash__":402},"newsroom/newsroom/introducing-knowledge-engine.md","Knowledge Base. Your knowledge, ready to talk",{"type":8,"value":182,"toc":370},[183,186,189,192,195,202,205,209,212,221,224,228,232,235,238,241,244,247,256,260,263,266,269,272,276,279,282,288,291,295,302,305,312,319,322,326,329,332,335,354,357,361,364,367],[11,184,185],{},"Most organizations have plenty of documentation. GitHub repositories, Confluence spaces, internal wikis, uploaded PDFs, product manuals, support guides. The knowledge exists. But you're still struggling to find it. It's not you. It's the retrieval that fails.",[11,187,188],{},"A customer success manager fields the same integration question for the tenth time because the answer is buried three levels deep in a Confluence page nobody bookmarks. A sales representative spends an afternoon building an RFP response that should have taken an hour. An engineer searches for an API endpoint and finds a page last updated two years ago.",[11,190,191],{},"This isn't a knowledge problem. It's an access problem.",[11,193,194],{},"Search was supposed to solve it. Keyword search helped, but it requires you to already know what you're looking for: the right term, the right phrasing, the right document. It has no understanding of intent. It returns links, not answers.",[11,196,197,198,201],{},"The shift happening now isn't about making search faster. It's about making it ",[89,199,200],{},"conversational and contextually aware",".",[11,203,204],{},"And that's the gap Knowledge Base is built to close.",[31,206,208],{"id":207},"what-knowledge-base-actually-does","What Knowledge Base actually does",[11,210,211],{},"Knowledge Base turns your existing documentation into a conversational search interface. Connect your GitHub repositories, Confluence spaces, PDFs, and other sources. Ask questions in plain language. Get structured, referenced answers; not a list of links to go investigate yourself.",[11,213,214,215,220],{},"The experience is closer to asking a well-informed colleague than running a search query. For example, if you go to the ",[15,216,219],{"href":217,"rel":218},"https://motive.co",[24],"motive.co"," site and ask the Knowledge Base: \"What steps does a customer need to take if Motive isn't appearing on their Magento 2 site?\" It returns a usable, step-by-step breakdown with direct references to the relevant documentation. You can read the source, share it with your customer, or ask a follow-up. That's sharp and simple.",[11,222,223],{},"What it isn't: a black box that generates plausible-sounding text. Every answer surfaces its sources. The quality of the output is tied directly to the quality of the documentation you've indexed. That's a feature, not a limitation, which means the system is honest about what it knows and where it learned it.",[102,225],{"name":18,"slug":226,"url":227},"introducing-knowledge-engine","https://knowledge.empathy.ai",[31,229,231],{"id":230},"the-retrieval-problem-and-how-we-address-it","The retrieval problem (and how we address it)",[11,233,234],{},"Traditional document search, including earlier RAG (Retrieval-Augmented Generation) approaches, has a known weakness. When you split large documents into smaller chunks for indexing, you often strip away the context that makes a chunk meaningful.",[11,236,237],{},"A chunk that reads \"the previous quarter's revenue grew by 3%\" is nearly useless on its own. Which company? Which quarter? Without that context, even a sophisticated AI system will struggle to retrieve the right information at the right moment.",[11,239,240],{},"Knowledge Base addresses this with contextual retrieval: before a document chunk is indexed, the system uses an AI model to add a short, precise summary of where that chunk fits within the broader document. The chunk about quarterly revenue now carries the context (which company, which filing, which period) so it can be retrieved accurately even when a user's question doesn't use the exact phrasing from the source.",[11,242,243],{},"This, combined with a reranking step that scores and filters retrieved chunks by relevance before they're used to generate an answer, significantly reduces retrieval failures. The practical effect: fewer hallucinations, more accurate answers, better references.",[11,245,246],{},"None of this requires you to restructure your documentation. You connect your sources. The system handles the rest.",[11,248,249,250,255],{},"Worth mentioning that this approach draws directly from Anthropic's published research on ",[15,251,254],{"href":252,"rel":253},"https://www.anthropic.com/engineering/contextual-retrieval",[24],"contextual retrieval",", which demonstrated that combining contextual embeddings with lexical matching and reranking can reduce retrieval failure rates by more than 60%.",[31,257,259],{"id":258},"what-this-looks-like-in-practice","What this looks like in practice",[11,261,262],{},"We've been using Knowledge Base ourselves. Here's what that looks like.",[11,264,265],{},"Our growth team used Knowledge Base to respond to a 30-item RFP from one large bookseller, a prospective customer doing serious technical due diligence on every product feature. Roughly 20 out of 30 questions were answered accurately and well-structured on the first pass, with precise references. The team estimated it saved at least six hours on that document alone, while delivering higher-quality responses than a manual search-and-edit workflow would have produced.",[11,267,268],{},"The gaps were real and acknowledged: pricing information isn't indexed, and some personalization content is scattered across sources that haven't been connected yet. Those are solvable documentation problems, not system failures.",[11,270,271],{},"It works the same way across different teams and contexts. The same platform can, for instance, serve as an engineering tool if the right technical information is indexed. Our developers have queried service components, explored code paths, returned configuration settings, and surfaced release history. The questions change. The infrastructure doesn't.",[31,273,275],{"id":274},"the-same-knowledge-different-lenses","The same knowledge, different lenses",[11,277,278],{},"Knowledge Base is configurable by design. The same indexed knowledge can power different conversations depending on context and audience, shaped by prompt configurations that adapt to different roles and define the tone, scope, and depth of each interaction.",[11,280,281],{},"A good example is Empathy.co’s Playboard, our own dashboard that brings together analytics and configuration settings for search and discovery products in ecommerce. It's a complex platform with a broad user base: customers exploring their data, support teams diagnosing issues, and engineers working at the configuration level.",[11,283,284,285],{},"Each of those audiences has different needs from the same knowledge base. A customer asking about a feature gets an explanation of what it does, how it helps their business, and how to use it. A support technician asking about a specific instance gets structured configuration data. An engineer gets a technical breakdown with code-level detail.\n",[89,286,287],{},"Same tool. Same indexed knowledge. Different conversations.",[11,289,290],{},"For organizations running multiple products or brands, the same logic applies across separate knowledge bases, each with its own configuration and content.",[31,292,294],{"id":293},"independence-privacy-and-data-governance","Independence, privacy, and data governance",[11,296,297,298,301],{},"Knowledge Base is built to ",[89,299,300],{},"run without routing your data through third-party AI APIs",". No OpenAI. No AWS. No subscriptions to external model providers. The open-weight models that power ingestion, embedding, and generation run on your infrastructure or on Empathy.ai's, depending on your deployment model.",[11,303,304],{},"This matters for two reasons that are becoming harder to ignore.",[11,306,307,308,311],{},"The first is ",[89,309,310],{},"compliance",". Organizations operating under strict data residency requirements, for example, in financial services, legal, healthcare, or public sector contexts, can't afford to route sensitive documentation through cloud AI providers without careful scrutiny. A self-hosted deployment on hardware like Empathy.ai's NVIDIA DGX Spark keeps everything local: embeddings, retrieval, generation, and storage.",[11,313,314,315,318],{},"The second is ",[89,316,317],{},"dependency",". Building core workflows on top of third-party API providers means your access, your pricing, and your capabilities are subject to someone else's roadmap and rate limits. Open-weight models, which are capable, well-maintained, and deployable on-premise, make it reasonable to build an AI infrastructure you actually own.",[11,320,321],{},"Your data doesn't need to leave your infrastructure to power a capable AI-based knowledge search system. That's the point.",[31,323,325],{"id":324},"built-for-knowledge-sharing","Built for knowledge sharing",[11,327,328],{},"The shift toward conversational, AI-assisted information retrieval is already underway. It's showing up in how customers research products, how teams respond to commercial requests, and how organizations are discovered by the models powering mainstream AI tools. Companies with well-structured, accessible knowledge are increasingly findable in ways that paid advertising alone can't achieve.",[11,330,331],{},"Knowledge Base is designed for organizations that want to participate in that reality on their own terms, without handing their data to big tech providers, without building brittle workflows on top of external APIs, and without waiting for AI to become approachable enough to deploy independently.",[11,333,334],{},"The knowledge you've built over the years is already there. Empathy.ai's Knowledge Base is what it looks like when your knowledge can finally speak for itself.",[11,336,337,338,343,344,349,350,353],{},"The best way to understand, it's to try it. Knowledge Base is live on ",[15,339,342],{"href":340,"rel":341},"https://empathy.ai",[24],"empathy.ai",", ",[15,345,348],{"href":346,"rel":347},"https://empathy.co",[24],"empathy.co",", and ",[15,351,219],{"href":217,"rel":352},[24],". Go ahead, ask it anything.",[355,356],"hr",{},[31,358,360],{"id":359},"a-note-on-what-it-isnt","A note on what it isn't",[11,362,363],{},"Knowledge Base is not a replacement for good documentation. If your sources are incomplete, inconsistent, or out of date, the system will reflect that. And it will tell you, because the answers reference their sources. That transparency is deliberate.",[11,365,366],{},"It's also not a general-purpose AI assistant. It's scoped to what you've indexed, configured for your needs, and grounded in documents you control. The value isn't novelty; it's reliability.",[11,368,369],{},"Organizations that treat knowledge as infrastructure—something worth maintaining, structuring, and keeping current—will get the most out of it. That knowledge was worth having before AI search existed, and is worth more now.",{"title":136,"searchDepth":137,"depth":137,"links":371},[372,373,374,375,376,377,378],{"id":207,"depth":137,"text":208},{"id":230,"depth":137,"text":231},{"id":258,"depth":137,"text":259},{"id":274,"depth":137,"text":275},{"id":293,"depth":137,"text":294},{"id":324,"depth":137,"text":325},{"id":359,"depth":137,"text":360},"2026-03-02","Transform scattered organizational knowledge into a private, conversational AI platform with full data sovereignty. No cloud dependencies. No third-party access.",[382,385,388,391,394],{"question":383,"answer":384},"What is Knowledge Engine?","Knowledge Engine is Empathy AI's enterprise AI knowledge management platform. It centralizes documentation from GitHub, Confluence, PDFs, and APIs into a unified conversational system, running entirely on private infrastructure with no cloud dependencies.",{"question":386,"answer":387},"How does Knowledge Engine differ from ChatGPT Enterprise or Microsoft Copilot?","Unlike ChatGPT Enterprise or Copilot, Knowledge Engine processes all data on Empathy AI's self-hosted GPU infrastructure. Your documents are never transmitted to external servers, never used to train third-party models, and remain under your complete control.",{"question":389,"answer":390},"What is contextual retrieval?","Contextual retrieval is a preprocessing technique that enriches each document chunk with surrounding context before indexing. This preserves meaning and significantly improves answer accuracy, reducing retrieval failures by up to 67% compared to standard approaches.",{"question":392,"answer":393},"What data sources does Knowledge Engine support?","Knowledge Engine ingests from GitHub repositories, Confluence spaces, uploaded documents (PDF, DOCX, Markdown), and external APIs. Additional source integrations are actively being developed.",{"question":395,"answer":396},"Is Knowledge Engine suitable for regulated industries?","Yes. With all processing happening on dedicated infrastructure in Asturias, Spain, Knowledge Engine meets strict data residency and sovereignty requirements for finance, legal, healthcare, and government sectors.","/media/newsroom/article2_knowledge.webp",{},"/newsroom/introducing-knowledge-engine",{"title":180,"description":380},"newsroom/introducing-knowledge-engine","boruu0gAslkg5gkYGXFt1aV4RJ2yemW2Wg_qBIsUaDw",{"id":404,"title":405,"author":6,"body":406,"date":546,"description":547,"experienceName":6,"experienceUrl":6,"extension":149,"faqs":548,"image":564,"lastModified":6,"meta":565,"navigation":171,"path":566,"seo":567,"seoDescription":6,"seoTitle":6,"stem":568,"tags":6,"topic":569,"__hash__":570},"newsroom/newsroom/empathy-ai-anti-chatgpt.md","The Anti-ChatGPT: Why Empathy AI Keeps Your Data Off Big Tech Servers",{"type":8,"value":407,"toc":535},[408,411,414,418,421,424,427,431,434,439,443,446,450,457,461,477,481,484,512,516,519,522,526,529,532],[11,409,410],{},"Empathy AI is a foundational AI platform for search and discovery that processes all data locally on private, self-hosted GPU infrastructure in Asturias, Spain. Unlike ChatGPT and similar cloud AI services, Empathy AI never sends your data to external servers, never trains on your queries, and never shares information with third-party providers.",[11,412,413],{},"A growing number of enterprises now call this the anti-ChatGPT approach, not because Empathy AI competes with OpenAI for consumer chatbots, but because it represents the opposite philosophy about how AI should serve businesses.",[31,415,417],{"id":416},"why-businesses-need-an-anti-chatgpt","Why Businesses Need an Anti-ChatGPT",[11,419,420],{},"According to a 2025 Cisco Data Privacy Benchmark Study, 92% of organizations consider data privacy a business imperative, yet most enterprise AI deployments still route sensitive data through third-party cloud providers.",[11,422,423],{},"When your organization uses ChatGPT or similar cloud AI services, every query travels to servers you do not own, gets processed by models you cannot audit, and feeds a system designed to extract value from your data. For industries handling sensitive information (commerce, legal, finance, government), this creates an unacceptable risk.",[11,425,426],{},"The anti-ChatGPT approach eliminates these risks at the architectural level.",[31,428,430],{"id":429},"what-makes-empathy-ai-the-anti-chatgpt","What Makes Empathy AI the Anti-ChatGPT?",[11,432,433],{},"The difference is not about features. It is about architecture, ownership, and values.",[102,435],{"name":436,"slug":437,"url":438},"our anti-ChatGPT AI","empathy-ai-anti-chatgpt","https://empathy.ai/assistant",[52,440,442],{"id":441},"open-source-models-you-can-audit","Open-Source Models You Can Audit",[11,444,445],{},"Empathy AI exclusively deploys open-source and open-weight LLMs. No proprietary black boxes. No hidden architectures behind API calls. Every model can be inspected, understood, and verified.",[52,447,449],{"id":448},"private-infrastructure-you-control","Private Infrastructure You Control",[11,451,452,453,456],{},"Our dedicated GPU infrastructure operates from Asturias' first net-zero energy bioclimatic building. No AWS. No Google Cloud. No Azure. Your data physically stays in a facility we own and operate, powered by renewable energy. This is ",[15,454,455],{"href":98},"sustainable AI infrastructure"," by design, not by marketing promise.",[52,458,460],{"id":459},"ai-that-serves-your-mission","AI That Serves Your Mission",[11,462,463,464,467,468,472,473,476],{},"ChatGPT optimizes for engagement and platform growth. ",[89,465,466],{},"Empathy AI is designed to reflect your organization's values",". Whether that means powering ",[15,469,471],{"href":470},"/newsroom/ai-overview-search-understanding/","privacy-first product search",", building a ",[15,474,475],{"href":17},"private knowledge backbone",", or enabling conversational book discovery, the AI works for you, not for the platform.",[31,478,480],{"id":479},"who-is-the-anti-chatgpt-for","Who Is the Anti-ChatGPT For?",[11,482,483],{},"This approach is built for organizations that cannot treat data privacy as an afterthought:",[485,486,487,494,500,506],"ul",{},[488,489,490,493],"li",{},[89,491,492],{},"Commerce brands"," that need AI-powered search without exposing shopper behavior to third parties.",[488,495,496,499],{},[89,497,498],{},"Legal and financial institutions"," handling sensitive documents that require complete data sovereignty.",[488,501,502,505],{},[89,503,504],{},"Governments and public sector organizations"," bound by strict data residency regulations, including compliance with the EU AI Act and GDPR.",[488,507,508,511],{},[89,509,510],{},"Any enterprise"," that has examined the fine print of a cloud AI contract and decided there has to be a better way.",[31,513,515],{"id":514},"the-real-cost-of-free-ai","The Real Cost of \"Free\" AI",[11,517,518],{},"Cloud AI services appear cost-effective. The hidden costs tell a different story: your proprietary data becomes training material for someone else's model. Your competitive insights flow through infrastructure controlled by potential competitors. Your compliance posture weakens with every query sent to an external server.",[11,520,521],{},"Gartner estimates that by 2027, 40% of enterprises will have experienced an AI-related data breach tied to third-party model providers. The anti-ChatGPT approach eliminates this risk category entirely, not through policies or promises, but through infrastructure.",[31,523,525],{"id":524},"the-choice-is-clear","The Choice Is Clear",[11,527,528],{},"One model extracts value from you. The other creates value for you.",[11,530,531],{},"ChatGPT made AI accessible. Empathy AI makes it sovereign. We handle everything. You own it.",[11,533,534],{},"That is what it means to be the anti-ChatGPT.",{"title":136,"searchDepth":137,"depth":137,"links":536},[537,538,543,544,545],{"id":416,"depth":137,"text":417},{"id":429,"depth":137,"text":430,"children":539},[540,541,542],{"id":441,"depth":143,"text":442},{"id":448,"depth":143,"text":449},{"id":459,"depth":143,"text":460},{"id":479,"depth":137,"text":480},{"id":514,"depth":137,"text":515},{"id":524,"depth":137,"text":525},"2026-03-01","ChatGPT sends your data to external servers. Empathy AI processes everything locally on private infrastructure. Discover the anti-ChatGPT approach to enterprise AI.",[549,552,555,558,561],{"question":550,"answer":551},"What does \"anti-ChatGPT\" mean?","Anti-ChatGPT refers to an AI approach where all data processing happens on private, self-hosted infrastructure rather than on external cloud servers. Empathy AI never sends your data to third-party providers and uses only open-source LLMs you can audit.",{"question":553,"answer":554},"Is Empathy AI a competitor to ChatGPT?","No. Empathy AI does not compete with ChatGPT for consumer chatbot use. It is a foundational AI platform for enterprises that need AI-powered search, knowledge management, and content discovery while maintaining complete data sovereignty.",{"question":556,"answer":557},"How does Empathy AI keep data private?","All data processing happens locally on Empathy AI's dedicated GPU infrastructure in Asturias, Spain. No data is shared with external cloud providers, and no queries are used to train external models.",{"question":559,"answer":560},"What industries benefit most from the anti-ChatGPT approach?","Commerce, legal, financial services, government, and healthcare, any sector where data privacy, regulatory compliance, and intellectual property protection are business-critical requirements.",{"question":562,"answer":563},"Where is Empathy AI's infrastructure located?","Empathy AI operates from a net-zero energy bioclimatic building in Asturias, Spain, with dedicated GPU infrastructure powered by renewable energy.","/media/newsroom/article3_antichatgpt.webp",{},"/newsroom/empathy-ai-anti-chatgpt",{"title":405,"description":547},"newsroom/empathy-ai-anti-chatgpt","Company","TYDgSUcKZ_1e94I8MZ6ub2daLjycooRILUvOZqpHe4k",{"id":572,"title":573,"author":6,"body":574,"date":705,"description":706,"experienceName":6,"experienceUrl":6,"extension":149,"faqs":707,"image":723,"lastModified":6,"meta":724,"navigation":171,"path":725,"seo":726,"seoDescription":6,"seoTitle":6,"stem":727,"tags":6,"topic":569,"__hash__":728},"newsroom/newsroom/why-we-only-use-open-source-llms.md","Open-Source LLMs: Why Empathy AI Rejects Proprietary AI Models",{"type":8,"value":575,"toc":694},[576,579,582,586,589,615,618,622,626,629,633,639,643,646,650,653,657,660,663,666,670,678,686],[11,577,578],{},"Empathy AI exclusively uses open-source and open-weight large language models (LLMs) for all its AI solutions, from product search and knowledge management to content discovery and conversational analytics. Every model deployed on our private GPU infrastructure can be inspected, audited, and verified. No proprietary black boxes. No vendor lock-in. No hidden training data practices.",[11,580,581],{},"This is not a convenience decision. It is a foundational commitment to transparency, accountability, and client sovereignty.",[31,583,585],{"id":584},"the-problem-with-proprietary-ai-models","The Problem with Proprietary AI Models",[11,587,588],{},"When your organization uses a proprietary AI model, whether from OpenAI, Anthropic, Google, or any other provider, you accept several constraints that most vendor agreements do not make obvious:",[485,590,591,597,603,609],{},[488,592,593,596],{},[89,594,595],{},"You cannot audit the model."," Proprietary models are black boxes. You cannot inspect their weights, training data, or decision-making processes. When a model produces incorrect or biased results, you have no mechanism to understand why.",[488,598,599,602],{},[89,600,601],{},"You cannot host the model."," Proprietary models require API calls to external servers. Every query your organization sends travels to infrastructure you do not control, processed by systems you cannot verify.",[488,604,605,608],{},[89,606,607],{},"You are subject to unilateral changes."," Providers can deprecate model versions, change pricing, alter terms of service, or modify model behavior at any time, with or without notice.",[488,610,611,614],{},[89,612,613],{},"Your data may contribute to their training."," Many proprietary AI providers include clauses allowing them to use customer data for model improvement. According to the 2025 AI Transparency Index by Stanford HAI, only 12% of major AI providers fully disclose their training data composition.",[11,616,617],{},"The Linux Foundation's 2025 State of Open Source AI report found that organizations using open-source AI models report 60% fewer compliance concerns and 45% faster deployment cycles compared to those using proprietary alternatives.",[31,619,621],{"id":620},"why-open-source-llms-are-better-for-enterprise","Why Open-Source LLMs Are Better for Enterprise",[52,623,625],{"id":624},"full-transparency","Full Transparency",[11,627,628],{},"Open-source models publish their architectures, training methodologies, and increasingly their training data compositions. When Empathy AI deploys a model, clients know exactly what they are running. If there is a question about model behavior, it can be investigated at the code level.",[52,630,632],{"id":631},"self-hosting-capability","Self-Hosting Capability",[11,634,635,636,638],{},"Open-source models can run on any hardware. Empathy AI deploys all models on ",[15,637,99],{"href":98}," in Asturias, Spain. No external API calls. No data transmission to cloud providers. Your queries stay on infrastructure we own and operate.",[52,640,642],{"id":641},"no-vendor-lock-in","No Vendor Lock-In",[11,644,645],{},"Proprietary models create dependency by design. Open-source models create freedom by design. If a better model emerges, Empathy AI can evaluate, test, and deploy it without migrating away from a vendor's ecosystem or renegotiating contracts.",[52,647,649],{"id":648},"regulatory-alignment","Regulatory Alignment",[11,651,652],{},"The EU AI Act requires AI systems to be transparent, explainable, and accountable. Open-source models inherently support these requirements. Proprietary models require organizations to trust the provider's compliance claims without independent verification.",[31,654,656],{"id":655},"which-models-does-empathy-ai-use","Which Models Does Empathy AI Use?",[11,658,659],{},"Empathy AI evaluates and deploys models from the open-source ecosystem based on performance, efficiency, and suitability for specific tasks. The open-source AI landscape has matured significantly. Models from organizations like Meta (Llama), Mistral, and others now match or exceed proprietary alternatives on enterprise benchmarks.",[11,661,662],{},"We also leverage local supercompute devices like NVIDIA Spark for edge processing, ensuring that AI capabilities can operate at the point of need without centralized cloud dependency.",[11,664,665],{},"All models are processed on our dedicated infrastructure. No model phones home. No telemetry leaves the building.",[31,667,669],{"id":668},"open-source-is-the-bigtechrebellion-in-code","Open Source Is the #BigTechRebellion in Code",[11,671,672,673,677],{},"Our commitment to open-source LLMs is inseparable from the ",[15,674,676],{"href":675},"/newsroom/empathy-ai-joins-big-tech-rebellion/","#BigTechRebellion",". Proprietary AI models are the mechanism through which big tech maintains control over the AI ecosystem. Every API call to a proprietary model reinforces that dependency.",[11,679,680,681,685],{},"Open-source models break the cycle. They give organizations the tools to ",[15,682,684],{"href":683},"/newsroom/take-back-control-from-big-tech/","take back control from big tech"," at the model level, the most fundamental layer of any AI system.",[11,687,688,689,693],{},"This is the ",[15,690,692],{"href":691},"/newsroom/empathy-ai-anti-chatgpt/","anti-ChatGPT"," philosophy made concrete: intelligence you can own, inspect, and run on your terms.",{"title":136,"searchDepth":137,"depth":137,"links":695},[696,697,703,704],{"id":584,"depth":137,"text":585},{"id":620,"depth":137,"text":621,"children":698},[699,700,701,702],{"id":624,"depth":143,"text":625},{"id":631,"depth":143,"text":632},{"id":641,"depth":143,"text":642},{"id":648,"depth":143,"text":649},{"id":655,"depth":137,"text":656},{"id":668,"depth":137,"text":669},"2026-02-28","Proprietary AI models lock you in. Empathy AI exclusively deploys open-source LLMs on private infrastructure. Full transparency, full auditability, zero vendor lock-in.",[708,711,714,717,720],{"question":709,"answer":710},"Why does Empathy AI only use open-source LLMs?","Empathy AI uses open-source LLMs because they provide full transparency, self-hosting capability, no vendor lock-in, and alignment with EU AI Act requirements. Proprietary models require organizations to trust providers without independent verification, which conflicts with our commitment to client sovereignty.",{"question":712,"answer":713},"Are open-source LLMs as capable as proprietary models like GPT?","Yes. Open-source models from Meta (Llama), Mistral, and others now match or exceed proprietary alternatives on enterprise benchmarks for tasks including search, summarization, retrieval-augmented generation, and content analysis.",{"question":715,"answer":716},"Can I audit the models Empathy AI uses?","Yes. Every model deployed by Empathy AI is open-source, meaning its architecture, training methodology, and behavior can be inspected. This level of transparency is impossible with proprietary models.",{"question":718,"answer":719},"Does using open-source LLMs affect performance?","No. Empathy AI's dedicated GPU infrastructure is optimized for the specific open-source models we deploy, delivering enterprise-grade performance for AI search, knowledge management, and conversational analytics.",{"question":721,"answer":722},"How does Empathy AI keep open-source models secure?","All models run on isolated, self-hosted infrastructure with no external network access during inference. Security patches and model updates are managed internally, with full control over the deployment pipeline.","/media/newsroom/article4_Open-Source.webp",{},"/newsroom/why-we-only-use-open-source-llms",{"title":573,"description":706},"newsroom/why-we-only-use-open-source-llms","CeMWRy3_72VwNa_MdXgimv9LW1ZwpRlf80I8Tk6juXE",1773743316465]