Go2Mars Research
humans.ai is committed to creating an integrated platform for AI-based creation and governance. On May 19, humans.ai announced the news that the product will be launched on the mainnet soon, and formulated a series of publicity and promotion activities around it. The initial focus of the project was to use artificial intelligence technology to generate or modify multimedia content such as images, audio, and video. But as time goes by, Humans.AI's ambitions have gradually developed, and its core vision is to make AI-type NFTs eventually become a digital extension of humans, or even replace humans to achieve digital immortality.
In 2007, the Svalbard gene bank, known as the "doomsday granary", was established on a small island about 1,000 kilometers from the North Pole. In this gene bank, about 100 million crop seeds from all over the world are stored in an underground warehouse at minus 18 degrees Celsius. Under the reflection of the Arctic icebergs, the pursuit of immortality and the sense of crisis of the end of the world collide in this gene bank to create a unique romance.
We all more or less believe that accurately replicating all our thoughts and memories through digital AI is a potential way of "digital immortality". But if we want to take it a step further and make it a reality, in addition to the memory and cognitive replication technology of digital AI, how to store our AI gene library is also a question that we have to consider - in what way, through what channels, and where to store it?
What if we use Web3 to build this artificial intelligence gene library?
Humans.AI Project Introduction
Since John McCarthy first proposed the concept of "artificial intelligence" in 1956, AI has been deeply rooted in people's minds, although it was more of an ideal concept at the time. After decades, the application possibilities of AI have been repeatedly confirmed, and now it has become one of the most popular technologies. Whether it is AI cloud computing in the past few years or AIGC applications represented by ChatGPT and Midjourney in recent years, they all demonstrate the value of AI.
However, in the new round of AI competition, in addition to technology giants, many participants regard small AI models and their databases as core competitiveness, further strengthening the concept of data sovereignty. But we should think: where should AI models and data be placed for real security? Thus, Humans.AI came into being.
Humans.AI is a blockchain-based content platform that focuses on the large-scale creation of AI applications on the chain, forming a multi-stakeholder ecosystem. It integrates AI tool libraries into creative studio suites, giving users more options when realizing their ideas. Users have the right to create and own their own digital portraits, which can be used by themselves and others to create countless digital assets. Synthetic media, AI applications, and other digital assets can all use blockchain technology to generate NFTs (non-fungible tokens) that cannot be forged.
Humans.ai provides users with an AI algorithm model library and a training and deployment environment (tradable), where developers can share their algorithm models, and other developers can also improve them after purchasing them. In addition, each developer will have a "digital gene" certification function, and users of the model need to obtain the author's permission through the "digital gene".
If we describe it in simpler and easier-to-understand language, we can understand that Humans.AI is an AI model blockchain content platform that can make AI models into NFTs for on-chain storage. Its core feature is to NFTize the "genetic" credentials that serve as human identity, such as digital images and voices, to ensure that every request forwarded to AI NFTs is verified as a real person, so that the scope of use of AI complies with the ethical standards of Humans.ai and a set of rules embedded in AI NFTs by its creators/owners.
How do AI NFTs work?
Imagine a researcher develops an algorithm that can take voice input and build a program that converts written text into audio that resembles the initial voice input - this is the underlying application logic of Humans.AI.
To ensure the security of the AI network, Humans.ai uses blockchain and a suite of other technologies such as cryptography and containers. The AI network is encapsulated in a container, a virtual machine-like structure that acts as a way to package an application so that it can run with isolated dependencies and its libraries and algorithms, so the application runs quickly and reliably from one computing environment to another. When the container receives compatible input, it generates an AI model.
AI NFTs are centered around the user’s digital DNA, and biometric data can be anything from voice, face, presence, gestures, etc. This information is encapsulated in AI NFTs and can serve as the basis for the development of future AI products such as synthetic media. For example, we can use multiple AI NFTs to create a talking avatar, a digital avatar that combines voice with face to create a digital character that can be trained to speak in multiple languages.
In addition to the biometric data it stores, AI NFTs also embed a set of rules that dictate how the information encapsulated within them can be used. In this way, we can prevent user-generated AI products from spreading harmful content, such as hate speech, politically incorrect information, or offensive content.
The process of creating an AI NFT in the Humans.ai ecosystem is called minting. Once an AI NFT is minted, it will contain two components:
▪️A descriptive component — biometric data and a set of governance rules outlined by the creator/owner of the AI NFT, as well as other data confirming the existence of the NFT on the blockchain
▪️Computational Components — A container structure that encapsulates an AI network. By employing a new type of zero-knowledge cryptography called zero-knowledge succinct non-interactive arguments of knowledge (zk-SNARKs), we are able to ensure that no entity can execute the AI component from the container without the digital signature of the AI NFT owner.
To facilitate user operations, Humans.AI has also built a rich infrastructure around trading markets, application stores, etc. Humans.ai plans to build an "AI Library" - deploy and train algorithms in a ready-made sales environment where users can share their creations and collaborate to improve their models.
Once the model is published, it will be automatically uploaded to Humans Studio. The data and algorithm models are encapsulated as AI entities and service products for commodity trading, and can be used by different users after being verified and authorized by "human credentials". This model means that certain forms of human intelligence - such as language communication ability, voice style, personality traits and other abstract elements can be traded and encapsulated into a performance entity.
Proof of Humanity (PoH): The Core of the AI Ecosystem
When artificial intelligence involves the application layer, the supervision of artificial intelligence systems becomes an issue that each of us has to face. Although "I, Robot" and "The Matrix" slightly exaggerate the dangers of artificial intelligence, if AI is allowed to develop freely, it will inevitably become a huge technological crisis.
It is based on this underlying logic with ethics as the main reference that Humans.AI launched the Proof of Humanity (PoH) consensus mechanism, which emphasizes the regulation of AI ethics on the basis of maintaining the fairness of the ecosystem. PoH is a complex and universal mechanism that ensures three key functions related to AI NFT: governance, consensus, and verification.
People can interact directly with the Humans.ai blockchain and PoH mechanism through a variety of smart interactive devices (smartphones, smartwatches, etc.) through dedicated applications. To ensure that the governance of AI NFTs is in line with human ethics, Humans.ai uses human proof to confirm that behind each request made to the AI model there are multiple digital signatures of real users, such as private keys and biometric data (fingerprints, heartbeats, retinal scans, facial recognition, etc.).
For example, let’s say a person named Jungle encapsulates her voice in an AI NFT. To prevent third parties from abusing her voice (a third party could be a media company or any person or entity that wants to generate content using an AI NFT), Jungle could engrave a set of rules in her AI NFT outlining how her biometric data can be used.
For example, Jungle can stipulate that his AI NFTs cannot be used to create obscene content, hate speech, or anything politically incorrect. The rules written inside the AI NFT can always be modified by its owner in case they wish to correct them or want to include additional rules.
The rules embedded in the AI NFT can also stipulate that only validators with a certain threshold of staked tokens can act as validators for the corresponding AI NFT, or how many validators are required to approve/reject a request. For their work, validators are rewarded in tokens, but if they consistently approve requests that deviate from the rules set in the AI NFT, they will be sanctioned and lose validator privileges.
Depending on their roles, human validators can be divided into two categories:
▪️Workers: Agents that analyze requests sent to the AI NFT to see if there are any discrepancies with the rules embedded in the AI NFT. Based on their analysis, Workers can choose to vote whether the request should be accepted or rejected.
▪️Block Producers: After the Workers vote on the request and give their approval, the Block Producers are responsible for actually validating the request. They are called Block Producers because after they give their final confirmation, they sign the request grouped in a block which is added to the blockchain to facilitate transparency and traceability.
By combining AI with blockchain technology, Human.AI empowers people to build and control their own AI, while ensuring that the development and use of AI products created in our ecosystem adhere to our ethical standards.
$HEART: The core of Humans.AI currency ecosystem
The ecological model also requires the flow of the token system to make the entire system truly operate. In Huam.AI, the system uses $HEART as the core currency.
According to the Humans.AI white paper, the current ecosystem mainly includes three token economic systems:
1. $HEART TOKENS: For governance and payment within the ecosystem
2.NON-FUNGIBLE TOKENS (NFTS): Ownership of algorithms, genomes, data, artificial intelligence, and applications
3. ERC20 TOKENS: Represents the governance and income rights of each NFT
As the native token of Humans.AI, the $HEART token enables anyone to participate in the governance process of the platform and facilitates the critical value flow within the platform. All fees collected by Humans.AI will be paid in $HEART, even if the payment originates from fiat currency or other acceptable cryptocurrencies.
One of the main functions of $HEART is to facilitate staking. Staking $HEART is required to perform key functions in the human ecosystem:
▪️Minting NFTs
▪️Account verification and transaction payment
▪️Create an app
▪️Adding AI to human-made operating systems
The four types of roles that contribute to the platform will automatically receive rewards based on the profit distribution algorithm. This idealized profit distribution method aims to ensure the rights and interests of data providers. It is a breakthrough in the Web2.0 data and algorithm monopoly model, returning the rights and interests of data and developers to the contributors themselves.
Governance is critical in the ecosystem because the community puts humans at the forefront of the AI revolution with the sole goal of aligning AI’s goals with ours. Governance is embedded in everything we do and it will play a core role in our ecosystem.
At this stage, Human.AI has 2 different levels of governance - protocol level and AI NFT level. Protocol level governance is well documented and will respect current governance standards and collaboration with communities and large projects such as Terra, Cosmos, Osmosis, etc.
Given that AI NFTs will generate economic value proportional to the utility of the NFT, owners of AI NFT DAO tokens will have a say in the $HEART pool that AI NFTs collect from accepted requests. For example, data providers of voice can set restrictions on what kind of voice casting requests can be accepted (such as only commercial videos), and human validators will check each request to comply with these restrictions. However, if data providers choose DAO style for product governance, the entire community will present a different picture under decentralized rules.
Summary
So, as a Web3+AI project with advanced concepts, has Humans.AI succeeded?
At least for now, they are still moving forward. Although the Humans.AI project is not perfect, the PoH mechanism is too ideal for the quantitative scoring of human ethics, and there are still some obstacles in the technical operation of encapsulating AI in the form of NFT. Even the drive of AI itself, according to the current stage of development, is not enough to achieve the effect of "digital twin".
But none of this matters. The greatest significance of humans.AI lies not in the project itself, but in what it tells us, the explorers:
When humans want to achieve digital immortality, Web3 will be the key to making thoughts immortal.
That's enough.
references:
[1]https://medium.com/humansdotai/heart-the-heartbeat-of-the-humans-ai-ecosystem-ed722f7c3c7e
[2]https://medium.com/humansdotai/humans-ai-presents-proof-of-human-d88bcec585dc
[3]https://medium.com/humansdotai/the-portrait-of-a-validator-in-the-blockchain-ecosystem-b04ca09a63bc
[4]https://medium.com/humansdotai/ai-on-blockchain-the-launchpad-of-humans-ai-ecosystem-b2f96eb8cc5a
[5]https://blog.humans.ai/blockchain-and-ai-how-they-can-work-together/
[6]https://blog.humans.ai/humans-synapse-enabling-token-transfers-between-two-distinct-blockchains-2/
[7]https://medium.com/humansdotai/the-humans-delegator-guidebook-457707ceb130
