Last month Binance published an official report dedicated to the fundamentals of decentralized AI.
Nowadays, most AI has been developed and is controlled by centralized Big Tech companies, and this gives them enormous power.
However, nowadays there are also large decentralized networks that work very well, and that could solve, perhaps even just partially, this problem.
Decentralized AI discussed in the latest Binance report
The decentralized artificial intelligence, also known as DeAI, is not solely composed of the widely used crypto AI agents on X, also because the latter often actually rely on centralized LLM (Large Language Model).
Instead, the DeAI movement aims to build networks and infrastructures to create AI models in a decentralized way from the outset.
AI is composed of three fundamental components, namely data, processing, and algorithms, and for years now, various DePIN (Decentralized Physical Infrastructure) projects have developed decentralized infrastructures that can be used as “building blocks” for DeAI projects.
The report from Binance states:
“As the race for artificial intelligence among the world’s superpowers begins to accelerate beyond individual control, distributed networks might start to attract attention as a means to consolidate resources and power related to artificial intelligence in a decentralized way”.
The Binance report on DeAI (Decentralized AI)
The report on DeAI published by Binance is a 21-page long pdf full of data.
Start by analyzing the current situation, which is dominated by centralized AI, and then delve into decentralized AI.
It addresses both the topic of AI agents and the fundamental basics of this great innovation, in order to provide a fairly complete picture of the situation, even from a technical point of view, but without becoming a technological treatise suitable only for the bull and bear technicians in the field.
Therefore, it is fundamentally within reach of all those who are interested in the issue, and it does not require extensive technical skills to be understood.
“`html
The conclusions
“`
After describing the current situation, the report reaches some interesting conclusions.
First of all, it reveals a growing intersection between blockchain technology and artificial intelligence, specifically in the form of DeAI.
However, it is noted how AI has long been at the center of public attention, while blockchain technology is experiencing a more gradual path towards acceptance by the mainstream.
They write:
“With the rapid advancement of AI progress, blockchain technology is emerging in parallel as a potential counterbalance to concerns about centralization. As the global AI race intensifies, blockchains and their distributed networks could provide promising avenues to mitigate the risks of centralized AI control”.
It is therefore still just a proposal and a hope, so much so that it does not even seem like an expectation yet.
“`html
The problemi dell’AI
“`
In this specific case, the key problem of AI is precisely that of being easily controllable by individual centralized entities.
In other words, since it is not publicly known exactly how the various AIs created by different companies operate in detail, one must in fact blindly trust them when turning to them to obtain, for example, information.
Sometimes it is possible to verify the truthfulness of that information at a later time, but it is not always possible to do so. Moreover, very few people choose to do it.
In this way, if the responses were guided by the centralized subject managing its operations, it might be impossible to be certain, and in some cases, it might even be extremely difficult to notice.
“`html
The DeAI solutions
“`
Transitioning from a centralized system to a decentralized one could simply eliminate control by a single centralized entity.
Furthermore, decentralized control would also have the advantage of being necessarily open source, meaning it can be publicly known in detail by everyone, since decentralized technologies require that all actors know in detail all or particular aspects of the protocol in use.
However, there are different levels of decentralization, as well as even a hybrid decentralization, therefore to achieve a situation realistically similar to the one just described, it would be necessary to use very well-decentralized systems.
In such case, an excellent starting point could be precisely blockchain technology, which in some cases (like Bitcoin) has not only proven to work very well but has also proven to be and remain highly decentralized for a long time.
Source: https://en.cryptonomist.ch/2025/03/21/the-binance-report-on-decentralized-ai-everything-you-need-to-know-about-the-new-trend/