Science fiction writer Isaac Asimov proposed the famous Three Laws of Robotics in his 1942 short story "Runaround":
First Law: A robot may not injure a human being or, through inaction, allow a human being to come to harm
Second Law: A robot must obey the orders given to it by human beings, except where such orders would conflict with the First Law
Third Law: A robot must protect its own existence as long as such protection does not conflict with the First or Second Law
Asimov's Three Laws are not actual technical specifications, but a literary creation that has profoundly influenced real-world robotics and AI ethics discussions, inspiring thoughts about AI safety, ethical design, and accountability.
In today's AI development, while the Three Laws are not directly adopted, similar principles (such as "human-centric" and "transparency") are often mentioned, especially in the context of Trustworthy AI. When discussing Trustworthy AI, some additional explanation is necessary.
Trustworthy AI aims to enable users to trust AI systems and confidently use them for decision-making or daily life while minimizing potential risks and negative impacts. How can this be achieved? If we apply the robot Three Laws, we might raise the following questions about AI development:
Safety: How can we ensure AI will not directly or indirectly harm humans?
Obedience: Should AI unconditionally obey human instructions?
Autonomy: How can AI maintain self-capability while being limited by human-defined boundaries?
To answer these three questions, AI design, development, and application must achieve reliability, safety, transparency, fairness, explainability, and privacy protection. The need to be both transparent and protect privacy is often criticized even in technological development, but this is the real requirement.
What can be done? Let AI continue advancing by using encryption technology to elevate Trustworthy AI, and even apply it on blockchain. Why? Blockchain naturally has openness and transparency, which inherently conflicts with AI's data sensitivity. So there's a small trick: if you see a project boasting about doing AI on blockchain, first check how they handle data encryption. If they can't handle that well, it's likely just riding the trend.
Encryption makes people's heads spin, with complex mathematics and technical terms that are incomprehensible when combined. Let me explain it in the most straightforward language, though I'm just a layman, and these encryption technologies truly require scientific experts to explain.
The most well-known is Zero-Knowledge Proof (ZK), once humorously mistranslated as "Zero Intelligence Proof". As a standalone technology, it is indeed a remarkable cryptographic achievement, primarily used to verify specific propositions by proving facts and outputting true or false results without revealing details.
The core principle is: do not disclose details.
For example, if you want to prove a wallet address belongs to you without revealing password or account details to any protocol or chain, you can use ZK to complete verification, ultimately providing a yes or no result.
Another recently discussed encryption technology is Fully Homomorphic Encryption (FHE). Another tongue-twisting term that is obscure and difficult to understand, but unavoidable since technical namers are often tech geeks who make understanding challenging. In the simplest language, it can be summarized as:
Performing calculations in an encrypted state and outputting encrypted results.
Is this comprehensible? Let me continue explaining. Traditional encryption methods (like AES or RSA) typically require decryption before data processing and re-encryption afterward. FHE's unique feature is supporting direct computation on ciphertext (encrypted data), with results consistent with encrypting the results of the same computation on plaintext (unencrypted data).
In other words, you give me a riddle, and I can work on the riddle's surface without knowing the answer, then output the result in riddle form, which only those knowing the solution can view.
This technology is now called the Holy Grail of encryption technologies because it perfectly solves the aforementioned problem of maintaining transparency while protecting privacy. The FHE concept was first proposed by Craig Gentry in 2009, and since then, academic and industrial sectors (like IBM and Microsoft) have continuously improved algorithms, such as those based on CKKS, BFV, or TFHE schemes.
Is there a blockchain project using Fully Homomorphic Encryption (FHE) and practicing Trustworthy AI? Indeed, that project is Mind Network. Has it issued tokens? Can it be farmed? Let's first discuss their basic situation.
Mind Network positions itself as infrastructure for on-chain intelligent agents, empowering developers to create fully encrypted blockchain networks. Binance Labs, Hashkey, Animoca Brands, Chainlink, and others have invested $12.5 million, and it has also received Ethereum Foundation support. Mind Network is the first FHE project integrated by DeepSeek, providing encrypted inference support for open-source models. Swarms has already collaborated with Mind Network on an AI multi-agent collaborative system, and ai16z, vana, and spore have also partnered.
Here, a technical term "HTTPZ" must be inserted.
We are already familiar with "http" and "https", where "http" was the basic protocol of early Web2 internet, all transmitted in plaintext with questionable security and privacy. Under the advocacy of companies like Google, "https" gradually replaced "http" as the universal protocol, but centralized privacy and security issues remain unresolved.
"HTTPZ" is a new protocol emerging from the FHE technology background, enabling data computation while maintaining encryption, achieving end-to-end secure transmission. AgenticWorld establishes AI Agents consensus based on this protocol.
The emergence of "HTTPZ" has given birth to an interesting topic: encryption sovereignty. If decentralized ledgers and decentralized intelligence merge, data citizens surviving in the "HTTPZ" era will be called CitizenZ.
The CitizenZ concept originates from Friedrich Hayek's free market ideology and principles proposed by Rees-Mogg and Davidson in "The Sovereign Individual". Hayek advocated minimizing external control and maximizing individual choice freedom. "The Sovereign Individual" further emphasized the importance of applying this freedom in the so-called "information age" (very similar to the intelligence age).
How to understand CitizenZ? It's simple: each person has absolute control over personal speech, data, assets, and other digital properties. These sovereignties must follow:
Removing intermediaries: Participation rights, like voting, do not require third-party mediation
Trustless security: System security is based on cryptography, not entities
Transparency: Completely verifiable processes based on blockchain, unaffected by tampering
Sovereign control: Individuals completely control basic rights like property, data, and voting
Taking citizen voting as an example, how would CitizenZ voting on blockchain and AI differ from current methods?
Verifiability: Using zero-knowledge proof to verify voting validity without revealing voter identity
Encrypted counting: Using homomorphic encryption for encrypted counting, ensuring voting fairness
Tamper-proof: Blockchain provides an immutable voting record, ensuring transparency
It's not difficult to understand why Mind Network has continuously received Ethereum Foundation support: as underlying technological logic gradually materializes, we can begin exploring deeper paradigms and orders, even implementing Hayek and Davidson's thoughts, and proposing a comprehensive philosophical foundation for the Agentic AI ecosystem.
Besides providing infrastructure for industry projects, Mind Network has first built an "Utopia" - AgenticWorld on BNB Chain and MindChain. This is a multi-chain intelligent agent economic system focused on training and collaboration, which can be simply understood as Mind Network creating a society of Agents (intelligent agents), even with schools and enterprises, enabling AI to learn and earn money in a one-stop growth process.
Here, users can create their own AI agents by staking tokens, and through the basic center, agents can continuously grow and earn rewards. When your agent reaches a certain stage, it can do tasks and work to earn money. If you are not satisfied with its performance, you can "kill" it and retrieve the staked assets (scary).
Did you notice? This is actually a self-running system with preset goals, based on the technologies mentioned above.
MindChain is a Rollup chain tailored for FHE verification, capable of processing large data and achieving fast settlement and transactions while maintaining high security. Through a reliable message transmission mechanism, MindChain can provide remote staking support for broader source chains, ensuring the credibility of the staking process.
The chain has now entered the tokenization phase, with airdrops already distributed. 11.71% of the total $FHE supply will be used for airdrops, with a minimum guarantee of at least 10 $FHE. Staking is now open, with the potential to earn up to 400% APY.
Yesterday, the token was listed on Binance Wallet, oversubscribed by more than 170 times. It is now tradable on Binance Alpha, Kraken, and other common exchanges.
