Author: OKG Research
Recently, OpenAI released a new model, Sora. This is a high-performance text-to-multimodal generation tool that supports generating detailed images and dynamic videos from text.
Compared to the video samples released earlier, this function can now be experienced by users. Registration has been suspended due to service overload. However, judging from the clips selected on the official website, the accuracy of the images and videos generated has been greatly improved. Some are even difficult to distinguish whether they are generated by AI. AI has become more concrete at this moment!
Pic Source: OpenAI Official Page
However, with the rapid development of AI technology, the threshold for generating false content has been greatly lowered, which has also brought more potential risks: AI-generated face-changing videos and deep fakes have begun to penetrate into daily life. Faced with this challenge, Web3 technology is bringing us new solutions.
Researchers from OpenAI, Harvard Society of Fellows, Microsoft, University of Oxford, a16z crypto and other companies have published a discussion on the value of a new tool to address this challenge: Personhood Credentials (PHCs). This digital credential enables users to prove that they are a real person without revealing any personal information.
The proliferation of AI-generated content: a crisis of trust
AI-generated realistic content (such as deep fakes), impersonation of others, and low-cost Sybil attacks are threatening the authenticity and security of online interactions.
Traditional countermeasures, such as CAPTCHA or anomaly detection, can no longer keep up with the pace of complex AI, while strict identity verification requires users to disclose personal information, undermining online privacy.
This dilemma has given rise to new demands for privacy protection and identity verification: How can we protect the anonymity of users while effectively verifying that online users are real humans, rather than fake identities generated by AI?
Personhood Credentials (PHCs): A new generation of identity verification tools
The article "Personhood Credentials: Artificial Intelligence and the Need for Privacy-Preserving Authenticity Verification Methods" proposes a very promising solution:
Personhood Credentials (PHCs) are digital credentials designed to enable users to prove to online services that they are real humans, not artificial intelligence (AI), without disclosing any personal information. These credentials can be issued by various trusted institutions (such as governments or other organizations), can be localized or global, and do not need to be based on biometrics.
Its emergence is expected to further cope with the future of widespread use of AI, where information or services generated by AI can be reasonably distinguished, and a "quasi-household registration system" between AI and humans can be established through personality proof.
In the online environment, PHCs can achieve real input and interaction from large numbers of real users, prevent fake accounts from disguising themselves as non-existent individuals, prevent robots from coordinating attacks to bypass platform rules and continue to abuse network services, and indicate that authorized AI Agents are agents of trusted users, not agents of malicious actors.
This will help reduce the proliferation of false information and identity fraud generated by AI, and ensure the credibility and security of the online ecosystem.
This solution does not need to rely on globally unified biometric verification, but can be issued locally by the government or other trusted institutions to adapt to different scenarios.
Pic Source:https://arxiv.org/pdf/2408.07892
Technical Implementation Path of PHCs
Currently, the concept of PHCs is mainly based on long-term research on anonymous credentials and "authentication" systems. Although the specific implementation plan is still under development, there are some related technologies that have laid the foundation for the implementation of PHCs.
Zero-knowledge proof (ZKP): PHCs can use zero-knowledge proof technology to allow users to prove that they are independent, real humans without disclosing specific identity information.
Decentralized identity (DID): Through the decentralized identity framework, users can create and manage their own digital identities without relying on central institutions.
Multi-party signatures and trust networks: Trusted institutions can jointly verify user identities through multi-party signatures to form a trust network.
Advantages of PHCs
Privacy protection:Users can verify authenticity without disclosing personal data, protecting digital privacy.
Enhanced trust:By preventing the spread of false identities and deep fake content, PHCs can establish a stronger foundation of trust for the digital society.
Non-biometric dependency:The verification process of PHCs does not rely on biometric data, avoiding related privacy and security issues.
Promote decentralized identity (DID) and platform development:PHCs can empower Web3 decentralized social platforms, allowing users to better control their personal identity and data.
Previously, there were Web3 projects in the industry that attempted decentralized identity, such as Worldcoin, launched by OpenAI co-founder Sam Altman in 2023, which attempts to provide biometric-driven decentralized identity and attempts to solve the problems of witch attacks and fake accounts.
However, the issue of iris collection and the subsequent security of related privacy data are widely controversial, but its large-scale trials have shown that users have a strong demand for simple, decentralized identity authentication methods.
Pic Source:The hardware device for Worldcoin
https://liamhorne.com/world-chain
Although PHCs have shown great potential, their deployment and implementation still face many challenges:
Security: Prevent PHCs systems from being abused or attacked, and ensure the security of data during transmission and storage.
Accessibility: Ensure that all types of users can easily access and use PHCs, especially those with low technical skills or limited resources.
Standardization: Unified technical standards need to be developed to ensure interoperability and scalability.
Possibility of combining AI with Web3
In decentralized social platforms, PHCs can be combined with blockchain technology. For example, when a user logs into a dApp, by verifying PHCs, the platform can confirm the authenticity of the user without collecting any sensitive data. This will provide effective support for dealing with deep fake content while preventing the proliferation of malicious robot accounts.
PHCs can also be used in decentralized identity (DID) systems to form synergies with areas such as copyright management of AI-generated content, anti-deep fakes, and community governance. For example, by binding PHCs, users can verify the authenticity of their identities in the Web3 ecosystem and ensure the fairness of governance decisions; and the combination of NFT and PHCs can provide traceability and copyright protection for AI-created content.
Through the combination of AI and PHCs, a trust incentive mechanism can also be built in digital economic activities. For example, in a decentralized work platform, employers can confirm the authenticity of freelancers through PHCs, and use AI analysis to match the most suitable collaborators. This model will effectively reduce the risk of fraud and improve collaboration efficiency.
From trust crisis to trust reconstruction
The significance of PHCs is not only to solve the current trust crisis, but also to lay the foundation of trust for the digital society. PHCs are expected to serve as the basis for international digital identity mutual recognition and reduce the risk of fraud in cross-border online activities. Through PHCs, humans can manage interactions with AI more effectively, thereby achieving a healthy development of human-machine collaboration.
By combining the transparency and decentralization of blockchain, PHCs provides a new imagination space for building a trusted digital identity and network ecology. PHCs will not only have an impact on the technical and commercial fields, but will also trigger in-depth discussions on privacy rights, digital identity and technical ethics. Policymakers, technical experts and standard organizations need to work with the public to promote the development and implementation of PHCs and create a safer and more reliable future for the digital society.