Science fiction writer Isaac Asimov famously proposed the Three Laws of Robotics in his 1942 short story Runaround:
First Law: A robot may not injure a human being, or, through inaction, bring harm to a human being.
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 may protect its own life unless it conflicts with the First or Second Law.
Asimovs three laws are not real technical specifications but a literary creation, but they have had a profound impact on real-world discussions on robotics and AI ethics, and have inspired thinking about AI safety, ethical design, and responsibility.
Although the three laws are not directly adopted in AI development today, similar principles (such as people-oriented and transparency) are often mentioned, especially in the context of trusted AI. Speaking of trusted AI, it is necessary to make some additional explanations first.
Trustworthy AI aims to enable users to trust AI systems and use them in decision-making or daily life with confidence, while minimizing potential risks and negative impacts. How is this done? If we apply the Three Laws of Robotics, we seem to be asking the following questions about AI development:
Safety: How to ensure that AI does not directly or indirectly harm humans?
Obedience: Should AI obey human instructions unconditionally?
Autonomy: How can AI maintain its own autonomy while also being constrained by the boundaries set by humans?
To answer these three questions, we need to ensure reliability, security, transparency, fairness, explainability, and privacy protection when designing, developing, and applying AI. The idea of both transparency and privacy protection is despised even in the field of technology development, but there is no other way because this is the real demand.
What to do? Let AI continue to move forward, use encryption technology to take trusted AI to the next level, and even apply it to blockchain. Why? Blockchain is naturally open and transparent, which in itself conflicts with the data sensitivity of AI. So here is a little trick. If you see a project bragging about doing AI on the blockchain, first see how it handles data encryption. If it cant handle this well, then it is likely to be a freeloader.
Encryption is a headache because it is so complicated. It is all mathematical formulas. I know every word but I have no idea what they mean when they are put together. Let me explain it in the most common language. Of course, I am only half-baked. The encryption technology here needs a real scientist to explain it.
The first one that everyone is familiar with is Zero-Knowledge Proof (ZK), which was once translated as zero IQ proof because of the backlash of zks. This technology alone is indeed a very powerful existence in cryptography. It is mainly used to verify specific propositions, prove facts without revealing details, and output true or false results.
The key is not to disclose details.
For example, I want to prove that a wallet address belongs to me, but I don’t want to tell any protocol or chain my password and account details. What should I do? At this time, you can use ZK to complete the verification, and finally give you a yes or no result.
Another encryption technology that has been hotly discussed recently is Fully Homomorphic Encryption (FHE). It is another extremely difficult word to pronounce, obscure and difficult to understand, but there is no way, the people who name the technology are often technology geeks, and it is really difficult to understand. In the simplest language, it can be summarized as follows:
Perform calculations in an encrypted state and output the encrypted results.
Is this understandable? I continue to explain. After the data is encrypted, if the data needs to be processed by traditional encryption methods (such as AES or RSA), it is usually decrypted first, processed, and then re-encrypted. The uniqueness of FHE is that it supports direct operations on ciphertext (encrypted data), and the operation results are consistent with the results obtained by performing the same operation on plaintext (unencrypted data) after encryption.
In other words, you give me a riddle, and I dont need to know the answer. I will just play around with your riddle, and then output the result in the form of a riddle that only people who know the answer can view.
This technology is now called the Holy Grail of cryptography because it perfectly solves the problem mentioned above, that is, how to protect privacy while ensuring transparency. The concept of FHE was first proposed by Craig Gentry in 2009. Since then, academia and industry (such as IBM and Microsoft) have continued to improve the algorithm, such as based on CKKS, BFV or TFHE schemes.
Is there any blockchain project that uses fully homomorphic encryption (FHE) and practices Trustworthy AI? There is indeed one, and that project is Mind Network. Has it issued a coin? Can I buy it? Let’s talk about their basic situation first.
Mind Network is positioned as the infrastructure of on-chain intelligent entities, enabling developers to realize a fully encrypted blockchain network. Binance Labs, Hashkey, Animoca Brands, Chainlink and others invested $12.5 million, and also received funding from the Ethereum Foundation. Mind Network is also the first FHE project integrated by DeepSeek, providing encrypted reasoning support for open source models. Swarms has cooperated with Mind Network to develop an AI multi-agent collaboration system, and ai16z, vana, and spore have also cooperated.
Here I would like to insert a technical term HTTPZ.
We are already familiar with http and https. http is the basic protocol of the early Web2 Internet, but it is all plain text transmission, and its security and privacy are worrying. Under the advocacy of large companies such as Google, https gradually replaced http as a universal protocol, but the centralized privacy and security issues have not yet been resolved.
HTTPZ is a new protocol developed under the background of FHE technology. It can calculate data while maintaining encryption and achieve end-to-end secure transmission. AgenticWrold is the consensus basis of AI Agents under this protocol.
The proposal of “HTTPZ” has given rise to an interesting topic: crypto sovereignty. If decentralized ledgers and decentralized intelligence are integrated, then the data citizens who survive in the “HTTPZ” era are called CitizenZ.
The concept of CitizenZ is derived from Friedrich Hayek’s ideas about free markets and the principles proposed by Rees-Mogg and Davidson in The Sovereign Individual. Hayek advocated minimizing external control and maximizing the freedom of individual choice. The Sovereign Individual further emphasizes the importance of applying this freedom in the so-called “Information Age” (which is very similar to the Smart Age).
How to understand CitizenZ? In fact, it is simple. It means that everyone has absolute control over personal speech, data, assets and other digital property. And these sovereign rights must follow:
Removing the middleman: Participatory rights, such as voting, do not require a third-party intermediary
Trustless security: System security is based on cryptography, not entities
Transparency: Fully verifiable process, based on blockchain, immune to tampering
Sovereign control: Individuals have full control over fundamental rights such as property, data, and voting
Take referendum as an example. If CitizenZ conducts voting based on blockchain and AI in the future, what changes will there be compared to now?
Verifiable: Use zero-knowledge proofs to verify vote validity without revealing the voter’s identity
Encrypted vote counting: Use homomorphic encryption for encrypted counting to ensure the fairness of voting
Tamper-proof: Blockchain provides an immutable voting record to ensure transparency
It is not difficult to understand why Mind Network has received continuous funding from the Ethereum Foundation: when the underlying technical logic is gradually implemented, we can begin to explore deeper paradigms and order, and even realize the ideas of Hayek and Davidson, and propose a complete philosophical foundation for the construction of the Agentic AI ecosystem.
In addition to providing infrastructure for industry projects, Mind Network has also built an ideal country on BNB Chain and MindChain - AgenticWorld. This is a multi-chain intelligent economic system centered on training and collaboration. In laymans terms, Mind Network has built an agent society, even schools and companies, to allow AI to grow from learning to making money in one stop.
Here, users can create their own AI agents by staking tokens. Through learning from the basic center, the agents can continue to grow and obtain rewards. When your agent grows to a certain stage, you can do tasks and work to earn money. If you are not satisfied with its performance, you can kill it and get back the pledged assets (scary).
Did you notice? In fact, this is a system that runs automatically after setting goals, and the underlying technology is all mentioned above.
MindChain is a Rollup chain tailored for FHE verification. It can handle big data and achieve fast settlement and transactions while maintaining high security. Through a reliable messaging mechanism, MindChain can provide remote staking support for a wider range of source chains to ensure the credibility of the staking process.
Now that the chain has entered the tokenization phase, the airdrop has been completed. 11.71% of the total $FHE supply will be used for airdrops, and the minimum wage for a minimum of 10 $FHE can be obtained. Staking is now open, and the highest APY can be obtained is 400%.
The new issuance was completed on Binance Wallet yesterday, with oversubscription of more than 170 times. It is now open for trading on Binance Alpha, Kraken and common exchanges.
