Artificial intelligence (AI) is a field of science that uses technology to create machines that can learn, reason, and act like humans.
The rise and roles of artificial intelligence (AI)
Artificial intelligence (AI) is a field of science that uses technology to create machines that can learn, reason, and act like humans.
Artificial intelligence (AI) has long been a fascination of its human, non-artificial counterpart. In literature and film, the concept of intelligent machines dates back to the mid- to late 1800s, with perhaps the earliest exploration by Samuel Butler in his work Darwin Among the Machines. Throughout the 20th century, AI was represented in books such as *I, Robot *and Neuromancer, and in TV/film projects like Star Trek. In 2001, the film A.I. Artificial Intelligence brought the concept of AI further into focus for the public.
Although AI remained purely fictional for many years, its real-life development began in the mid-1900s. Early efforts were bare bones, but they were considered significant breakthroughs at the time. However, the blossoming of computing capabilities in the 1980s and 1990s primed a flourish of successes in the 2000s. This is when machine learning and “general” AI began development, in earnest.
In November 2022, ChatGPT showed how AI could be useful on a day-to-day basis. As a publicly accessible large language model (LLM), its launch exhibited the advancement of AI technology and showed the world that an AI chatbot model could approximate interactions with a human. It saw adoption by more than 1 million people in its first five days and claimed more than 100 million users in its first two months. Over the following year, other LLMs were released by competitors including Bard/Gemini (Google) and Claude (Anthropic).
While LLMs have arguably drawn the most public attention, other forms of AI have readily joined them in rapid succession. Some of these can generate pictures or videos from text input, others help users make music or convert text into human-like voices, and still others aid in photo editing (e.g. removing unwanted objects). Industry implementations include analysis of big data, autonomous vehicles, and improving research efforts through neural networks.
Decentralization of AI through blockchain
Most of the existing AI platforms are run by companies acting in isolation. The aforementioned LLMs, for instance, are entirely owned and operated by their respective enterprises: OpenAI, Google, and Anthropic. These companies train their AI models using a trove of data that is scoured from sources that are, in many cases, copyrighted. This brings into question both the ethics of AI training and the transparency of each model. Users do not own their data, nor do they know exactly how the model was built.
Decentralized AI aims to bring the concepts introduced by blockchain and cryptocurrency technologies into the AI space. Just as decentralized finance (DeFi) alters the basics of traditional finance, decentralized AI hopes to improve the existing system. Distributing data provision and processing across a broad network of nodes can improve security and make models fairer and more transparent (and thus trustless). Furthermore, built-in economic incentives—a familiar concept to anyone in the crypto space—promise financial reasons to improve and maintain these systems. One might imagine a LLM which is trained based on users’ data, where the users are compensated fairly for their participation, and the model is therefore better understood and more effective.
Early examples of blockchain-based decentralized AI
Fetch.ai – One of the most promising uses for AI is performing certain kinds of work without human input. Fetch.ai provides AI-powered “agents” that are registered to its blockchain using smart contracts. Agents can autonomously complete tasks for owners. Its native cryptocurrency is FET.
Ocean Protocol – Ocean is described as a “decentralized data exchange protocol,” which provides a marketplace for buying and selling data. Users can provide private data to consumers (e.g., for market research) while maintaining their privacy and receiving compensation in return.
The future of decentralized AI essentials
Artificial intelligence became a worldwide phenomenon in 2022, but its functionality is not limited to the large language models (LLMs) that first caught the public’s eye.
AI models have so far mostly been centralized—owned and operated by individual corporations using opaque models and sources of data.
Decentralized AI projects like Fetch.ai and Ocean Protocol aim to democratize AI tools through blockchain technology to create data marketplaces, create AI-powered agents, and build other services.