“Language is the underlying operating system of human civilization, and LingoAI’s LingoPod, as a decentralized AI physical infrastructure network (AI+DePIN), is the driver for building the Tower of Babel.” Initially, God confused human language, but now AI breaks down language barriers, enabling people to communicate seamlessly across languages.

The industry networking event “The Moment of DePIN’s Mass Adoption,” organized by the renowned Chinese Web3 community 1783DAO, was held at Central Exchange Square, Hong Kong. One of the most popular and professional side events during BitcoinAsia, this event attracted over 500 participants.

At the event, LingoAI’s founder Una Wang discussed the Tower of Babel, one of the most iconic legends in human culture and religious history, concerning the future of people’s AI. The Tower of Babel is not only a biblical story about heaven and hell, humans and God, Satan and God but also a profound reflection on the origin and evolution of human language. In ancient Hebrew, “Babel” means “confusion,” revealing the mysterious origins of human language confusion. In the field of artificial intelligence, the Tower of Babel has always been synonymous with the utopian ideal of achieving barrier-free communication across languages through AI.

Initially, humans spoke the same language and communicated without obstacles, enabling smooth cooperation and understanding. This led to the grand idea of building a tower to reach heaven and directly communicate with God. God saw this and was worried about human actions, discovering Satan’s hidden intentions behind the people. Instead of destroying the tower or punishing the builders, God chose a more subtle and long-term method to correct human development — confusing their language. Suddenly, the builders, who could previously communicate fluently, could no longer understand each other, forcing the cessation of the Tower of Babel’s construction.

The confusion of languages caused great bewilderment and panic. People began forming small groups with others who spoke the same language. This change in language not only affected writing and pronunciation but also deeply influenced people’s thinking, value, and worldviews. Over time, these language-based groups developed distinct cultures and customs. They “multiplied and filled the earth,” settling in different corners of the world, each with unique cultures, beliefs, and ways of life. Thus, when God made the builders of Babel speak different languages, humanity developed diverse cultures, ideologies, and religions. However, this also led to more discord, conflicts, and even wars. Therefore, the story of the Tower of Babel is not just about a tall tower and language confusion; it touches on the core values and dilemmas of human society.

By the 20th century, the famous philosopher Wittgenstein said, “The limits of my language mean the limits of my world.” In the field of artificial intelligence, which began in 1956, the Tower of Babel has always been a symbol of humanity’s hope to achieve barrier-free communication through AI. In 2017, a paper titled “Attention is All You Need” introduced the Transformer architecture to solve machine translation problems. The authors may not have realized what they had invented. The greatest value of Transformer is changing how machines understand human natural language. After OpenAI launched ChatGPT, the boundaries of language seem to be breaking down.

LingoAI’s AI Agent, LingoPod, empowered by the Transformer large language model, is opening new ways for communication between humans and machines and among people of different languages. It provides real-time translation for different languages and cultural groups, helping them overcome language barriers and promote deeper understanding, cooperation, and exchange. This fundamentally changes the relationship between humans and AI, as it is no longer seen as a smart earphone or translation tool but as a means for genuine intellectual exchange and dialogue, developing a new partnership between silicon-based and carbon-based life.

Of course, reaching this level requires further evolution and training in large language models. Models capable of cross-lingual knowledge transfer need more corpora. There are over 7,000 languages in this world, but except for major languages like English and Chinese, most are low resource languages, resulting in poor reasoning and translation performance for large language models, especially in linguistically diverse regions like Indonesia. Currently, most high-quality speech recognition systems can only recognize about 100 languages. The lack of corpora presents a challenge for AI proliferation in developing countries.

To address these issues, LingoAI launched DePIN x AI device LingoPod, the world’s first Web3.0 AI Agent. It is not only a multi-language real-time translation tool that uses voice for human-machine interaction as an AI Agent, helping users set schedules, send emails, and organize meetings through natural language interaction. It also builds a decentralized corpus physical infrastructure network based on edge computing and federated learning. Users can contribute speech corpora through LingoPod to earn Lingo tokens, known as corpus mining.

In a global corpus DePIN network built from the bottom up by 100,000 LingoPods, a sub-network of 100 LingoPod nodes can accumulate 50 hours of diverse speech data in a particular one language within just 30 minutes. Compared to centralized corpus collection methods, this Web3.0 approach is more efficient and decentralized (without middlemen taking a cut). Users receive fair token rewards after their data contributions are validated by ReviewDAO.

LingoAI’s vision is to connect the world through DePIN and AI technology, enabling large models to efficiently support all human languages and let everyone embrace AI.

God confused human languages, but AI is breaking language barriers, allowing people to communicate freely across languages again. Yuval Harari, author of “Sapiens, A Brief History of Humankind” marvels, “AI has hacked the operating system of human civilization.” Does this mean that humanity, in creating AI, has violated God’s will?

Una explained that in the Bible, God confused human language initially to ensure humans would “multiply and fill the earth” (distributed and decentralized) because he saw Satan’s intention to place the “all-seeing eye” (not God’s eye) atop the Tower of Babel, watching over all and instilling fear and panic. Now, centralized AI will undoubtedly be controlled by a few behind AI giants, with personal, corporate, organizational, and even national security and privacy monitored and manipulated by the “all-seeing eye.” The supreme AI capabilities would inevitably lead to Satan-like evil. Only by returning data ownership to users and protecting data privacy through Web3.0 can AI develop sustainably and decentralized. Such a Tower of Babel is what God approves of and what people aspire to — a tower to heaven.

LingoAI’s initial intent is to build an AI for all humanity and has registered the domain PeopleAI.io. If everyone can gradually own their data through Web3.0 protocols (MetaLife.Social and SOLID by the father of the World Wide Web), combined with edge-side AI large models, they will create AI Agents that serve each person, enterprise, organization, and government privately. Therefore, LingoAI defined the concept of a Web3.0 AI Agent.

Bill Gates predicted: that GPT-5 is not the future; AI Agents are the real game changers. Una believes that centralized AI Agents are the “all-seeing eye,” spies placed by AI supercomputing platforms in user-end devices. Only decentralized Web3.0 AI Agent combined with centralized super AI represent the full panorama of future human AI.

In her keynote speech, Una introduced LingoAI’s overall architecture, including the protocol layer, model layer, and application layer. The protocol layer adopts semantic web, knowledge graph, and peer-to-peer communication, the model layer is based on different large models for personalized training, and the application layer provides private model deployment and data processing services. Every company hopes to have its own large model, but meanwhile face challenges on private deployment due to high GPU costs and reluctance to use data for centralized AI model training. However, through the MetaGraph technology stack, data is released, returning data ownership to users while enabling data interconnectivity between enterprises and individuals, aiming to provide personalized and professional AI models for each person, government, and enterprise.

Una, certified in data privacy and technology from Harvard University, explained that LingoPod uses the most advanced SOTA models and neural network translation technology, capable of translating 454 languages and recognizing 100 spoken languages. LingoAI protects user data privacy using the MetaLife protocol, allowing federated learning and edge computing without data leaving the LingoPod device. Users need not worry about data security and privacy issues. LingoPod is now available for pre-sale.

Talking about the founding of LingoAI in Singapore, Una thanked one person: Sopnendu Mohanty, Chief FinTech Officer of the Monetary Authority of Singapore (MAS). Mr. Mohanty carefully explained to Una the need for privatized deployment of AI large language models to protect confidentiality and privacy and the government’s necessity for translated documents to be checked. Una and her partners were discussing combining Web3.0 and AI large language models to build a Tower of Babel, enabling barrier-free communication. Mr. Mohanty’s reminder gave the project a clear B2B model. For LingoAI’s first Web3.0 AI Agent LingoPod for the B2C market, Mr. Mohanty places great hopes. He said it could be promoted at the Singapore FinTech Festival hosted by MAS in November, as thousands of fintech professionals attending the event would need this multi-language real-time translation tool.