Proof of Identity for AI Creations @OpenGradient, @idOS_network, @opensea As creations utilizing AI technology rapidly spread, confusion and plagiarism controversies surrounding who created them have repeatedly arisen. In particular, in an environment where images or videos generated by AI are circulated in the form of NFTs, it often becomes difficult to clearly distinguish the responsible parties using only existing copyright concepts. Against this backdrop, the identity verification structure of AI creations combined with OpenGradient, idOS, and OpenSea is gaining attention for technically answering the question of to whom responsibility should be attributed, rather than questioning the creator of the work. The starting point of this structure is to prove the actual process through which AI generates its outputs. OpenGradient records the execution process of AI models on the blockchain through an AI computation architecture called HACA. In this process, inference nodes perform actual calculations, while full nodes verify whether the calculations used the specified model and parameters, and storage nodes and data nodes maintain the integrity of the model and input data. As a result, a unique transaction record is left for each AI output, indicating which model was used and under what conditions it was generated. This means that an AI output is not just a simple file but a product with a verifiable creation history. However, proving the generation process alone is not sufficient. It is necessary to connect who operated the AI and who is responsible for the output. This role is fulfilled by the decentralized identity system, idOS. idOS provides a reusable credential form after verifying the identity of individuals or organizations once through a passporting and data ingestion structure. This allows creators to prove that they are verified entities without disclosing their real names, and they can limit the disclosure of information only when necessary. This structure operates as a technical compromise to maintain both anonymity and accountability. The link that combines AI generation records and human identities is handled by the Ethereum Attestation Service (EAS). The AI inference results generated by OpenGradient are assigned a unique hash, which is then combined with the hash of the identity credential issued by idOS to be recorded as a single attestation. This attestation can be stored in either on-chain or off-chain forms and can ultimately be included in NFT metadata. As a result, NFT buyers can cryptographically verify which AI model generated the work and which human entity is responsible for the output. This information can be directly utilized when registering NFTs on OpenSea. Since OpenSea already supports an IPFS-based metadata structure, it is possible to include the EAS attestation identifier as a property of the metadata. While this does not replace existing transaction volume-based authentication or manual reporting procedures, it provides an additional layer of trust. Particularly in cases where plagiarism allegations arise, this differs from the existing DMCA process, which can take days, as it allows for immediate verification of the creation history and identity connection. This structure also aligns with the legal environment as of 2025. In the United States, the Thaler v. Perlmutter ruling has clearly established that copyright belongs only to human authors, and the EU's AI law emphasizes the responsibility of human operators for the results of AI systems. In this context, the method of clearly recording the human responsible for operating and utilizing AI, while not recognizing AI itself as a copyright subject, helps bridge the gap between institutional demands and technological realities. Of course, limitations exist. It is technically challenging to completely block actions like proxy minting, where a verified user provides a signature on behalf of another. Additionally, there remains a tension between privacy protection and dispute resolution. Furthermore, this structure does not verify whether the training data for the AI model was lawful, leaving copyright issues at the model level as a separate challenge. Nevertheless, the identity verification structure connects the generation process of AI creations and the responsible parties into a single verifiable record, transforming the issue of AI plagiarism controversies from post-dispute to pre-verification. Ultimately, the model combining OpenGradient's computational proof, idOS's decentralized identity, EAS's attestation, and OpenSea's NFT distribution structure is establishing itself as a realistic way to technically address trust issues surrounding AI creations. This shifts the discussion surrounding works created by AI from ambiguity of the creator to clarity of responsibility, providing a foundation for verifiable copyright management in the digital creation environment.