#2 Reputation in the Age of AI

By LJW, Jack Butcher and Sven Wellmann

In this piece, we explore how the concept of reputation is changing with the introduction of generative AI. As AI-generated content becomes increasingly indistinguishable from human-created work, traditional notions of credibility and trust are being fundamentally challenged. We examine how reputation currently functions in the Internet era on social media platforms and highlight how existing issues are dramatically amplified by AI technologies. In response to these challenges, we explore the potential of crypto-based systems for reputation management. Could blockchain technology provide a solution to verifying identity and credibility in an AI-dominated digital landscape?

This piece was written with Jack Butcher, who is a leading digital artist. Before Web3, he founded Visualize Value, a community of Internet creators. In Web3, he created Checks and Opepen, two of the more popular digital art projects using blockchain as a canvas to comment on the changing nature of verification on the Internet (Checks) and how network effects can accrue through remixing a common canvas (Opepen). His artworks have also been featured and auctioned at Sotheby's and Christies.

As a philosopher of the Internet, Jack has been publicly exploring how technologies like AI and blockchain are transforming the digital landscape.

The introduction of Generative AI has democratized content creation, blurring the lines between amateur and professional creators. These technologies enable anyone to become a prolific creator, producing high-quality content at an unprecedented scale with near zero marginal costs. In the era of AI, it’s not the craft that creators need to master, it’s the prompt. A well-instructed AI model brings centuries of craftsmanship to the fingertips of countless individuals. By dramatically lowering the barriers of entry to creating, AI has unleashed a flood  of new creators and artists, bringing with it a fresh set of challenges – chief among them, the question of authenticity.

The proliferation of AI-generated content has dramatically shifted our perception of authenticity on the Internet. With infinite content, we are transitioning from a world where content was assumed to be real and genuine unless proven otherwise, to one where we must assume the content is either fake or AI generated, unless proven otherwise. This new paradigm forces us to question the veracity of all digital content, altering how we consume and interpret information in the AI era.

This is largely part of the reason creators have been one of the first groups to push back against AI. Creators are concerned that if an AI can produce work in their style, the value of their work will diminish. If you have a distinct visual style that is being used by AI, and the AI naturally extends what you are able to produce, as a creator you are probably frustrated that all the hard work it took you to get here is replicable in a click or a prompt. This has led artists and creators down a path of lobbying for monetization systems, citing that if an AI model uses an artist's image, the model creators should compensate the artist for that image. While this is a reasonable argument, the challenge is then attribution and answering what constitutes an image that is in the style of the artist in question. Additionally, in an AI model with trillions of data points in its training data, even if the attribution can be solved, the payouts might not be that meaningful without the ability to pinpoint the value contributed to the output of these models. What is clear is that attribution and authenticity in the age of AI has complex implications that need to be addressed. 

The debate around AI-generated works and value capture can be reframed to focus on how the concept of reputation changes in the age of AI.

If a well-known, respected artist and someone entirely unknown were to create nearly identical artwork, what is the difference between the two pieces? Given their likeness, would art critics judge them equally? Would both artists be able to command the same price in an auction? 

The answer is (likely) no. 

That’s because the artwork itself, while being the assumed focal point of appreciation,  is only one part among many that make an artwork notable. The commodification of creation by AI further erodes the value of the image by itself – in a world of infinite content it’s no longer the rare thing. A much bigger variable in this equation is the reputation of the creator. The challenge for creators is now not just creating quality content, they also have to worry about building a reputation in a world where AI can replicate their styles and techniques instantly. Reputation influences their ability to capture value and is the difference between a penny postcard and a million-dollar masterpiece.

As more and more of our lives move from the physical world to the digital world, so does our reputation. For any up-and-coming creator, having a digital presence becomes table stakes. Today, reputation on the Internet is largely arbitrated by social media platforms. The legitimacy of a piece of content is a function of platform-based reputation and the social consensus that reputation can command. However, the existing paradigm creates a few problems for creators accruing reputation for their work:

• Walled Gardens: Imagine you post an original image on X, where you have a large following. Your act of posting implicitly claims authorship. However, someone else could easily repost that same image on LinkedIn - where you lack a significant presence but they have a substantial following - and falsely claim it as their own creation. In this scenario, who decides what's perceived as genuine, and what is seen as fake? The social media platforms decide what is "real" or legitimate on their platform, but the same piece of content can be labeled differently on other platforms. This fragmentation also creates significant inefficiencies for creators. Creators must build and maintain separate reputations on each platform, duplicating efforts across different platforms. Moreover, if a platform shuts down or changes its algorithms or policies, a creator's hard-earned reputation can vanish overnight.

• Small Creators vs Large Creators: Consider this scenario: a budding creator with minimal online presence produces a compelling piece of content. A more established account then appropriates and shares it as their own. Which post is perceived as genuine? The larger account could also easily claim that they created it. In this instance, what recourse does the small creator (the original creator of the piece of content in question) have? They have limited options since there is no tamper-proof way on social media platforms to prove that the small creator created the image, and pursuing the offline route via the legal system is arduous and cost prohibitive for many creators. 

• Attestation to AI-Created Content: It's increasingly common for AI to remix a piece of content to create a derivative work. In this instance, should the original creator get any credit for creating the precursor content? Current social media platforms lack robust mechanisms to properly attribute or credit the original creator of the precursor content in AI-generated derivatives. These platforms were not designed with generative AI in mind, and their existing attribution systems are ill-equipped to handle the complexity of AI remixes.

A critical challenge in the current system is what we might call the “attribution chasm.” Creators who are willing to put in the work of producing hundreds or thousands of pieces, but lack proper attribution mechanisms, cannot accrue the reputation and build a network around their work required to be successful given the problems highlighted above. 

These limitations call for a new approach to digital identity, reputation, content attribution, and provenance. This new system also needs to withstand the challenges posed by AI-generated content, which are exacerbated by the current structure of social platforms, to provide creators with more control over their work and offer interoperability across various platforms and ecosystems.

For the challenges creators face, crypto's primary value proposition is a portable reputation that is formed based on the history and provenance data of a creator's wallet. On-chain actions and data establish the foundation for a new type of reputation system. In this system, signing a piece of content serves as verification, with the signature becoming a verified entity in its own right. Crypto enables this form of reputation building based on the principle that each user controls their own private keys, creating a unique chain of provenance data. Similar to how a social media account accrues reputation based on posts and interactions, wallets accrue reputation for actions performed on-chain. As a user accumulates more reputation, their signature carries more weight.

This system differs from traditional social media platforms, where followers can be purchased, and bot accounts can inflate numbers. While both crypto wallets and social media accounts start with minimal reputation, crypto enables tamper-proof attestation that operates across multiple applications due to its interoperable and transparent nature.

Imagine a fledgling digital artist named Jack who creates unique digital images. In the current social media landscape, Jack might post their work on Instagram and Twitter, building separate followings on each platform. If an AI system generates artwork in Jack's style, or if another user reposts Jack's work without credit, it can be challenging to prove original authorship across these disparate platforms.

Now, consider a crypto-based system:

1. Jack signs each piece of artwork with their private key and publishes each piece on a blockchain before posting it anywhere online. This creates an immutable timestamp of Jack's authorship.

2. When Jack posts the artwork on various platforms, they include a link to the blockchain record. This link serves as proof of authorship, regardless of where the art appears.

3. If an AI system generates similar artwork, it won't have Jack's signature. Users can check the blockchain to distinguish between Jack's original work and AI-generated imitations.

4. If another user reposts Jack's work without credit, Jack can prove authorship by referencing the blockchain record, which predates the repost.

5. As Jack continues to create and sign artwork, his wallet builds a verifiable history of creation. This history forms the foundation of Jack's reputation, which is portable across any platform that can read the blockchain.

6. Other users can "cosign" Jack's work by interacting with it on-chain (e.g. purchasing, commenting, or sharing). These interactions further build Jack's reputation and create a network effect around his work.

This system solves the problems of AI imitation and platform-specific reputation by providing a universal, tamper-proof record of creation and interaction that exists independently of any single platform. 

An astute skeptic will call out that the benefits of the above system will only work if enough people adopt this system as the default standard. Said another way, until a crypto based system generates more network effects and value for its users than existing platforms, the average user might still default to existing social media platforms. This is a fair point and is the adoption problem that all emerging technologies must go through. However, we are seeing an increasing amount of creators wanting a different system, especially in light of AI (e.g.Hollywood strikes and artists seeking protection against AI).

An even more astute skeptic might argue, "What prevents someone from signing work they didn't create?"  

While it's true that anyone can sign any piece of work, the power of crypto-based systems lies in the network effects and the accumulation of verifiable actions over time. In this scenario, social consensus and network effects enabled by a crypto based system can help arbitrate.

Even if a bad actor manages to register an artist’s work on blockchain before the artist does, if the artist's network recognizes them as the true creator and continues to engage with and build upon their work, these on-chain actions (such as which works collectors hold, which pieces of media contributors remix, and the verifiable provenance data of the artist's broader body of work) serve as powerful signals of authenticity. These collective actions and interactions, recorded immutably on the blockchain, create a complex web of evidence that is exceedingly difficult for a fraudulent actor to fabricate or maintain over time. As crypto technology matures, we can also expect new on-chain primitives to emerge that help arbitrate these issues (e.g. dispute resolution modules built directly into blockchain protocols or smart contracts.) 

Crypto is fundamentally about network formation. By putting their work on-chain, creators can build a repository that doesn't depend on a single platform. Over time, the network and social consensus that forms around a piece of media becomes more important than the media itself. It’s easy to copy the work, but it’s hard to replicate the network.

Aside from helping artists secure their work and create an interoperable reputation, crypto based systems also enable creators to become antifragile to AI and remixing. As AI generates unlimited content, an artist's crypto-based reputation becomes the scarce and valuable resource. AI might generate similar images, but it cannot generate the rich tapestry of interactions, history, and community surrounding an artist's body of work in a crypto-based ecosystem. And it cannot fabricate years of consistent creation and innovation, recorded on a blockchain. This reputation, built through consistent creation and community engagement, becomes more valuable than any single piece of work. 

The individual nodes – collectors, collaborators, and fans – that interact with the creator’s work on-chain add value to the creator's network. These network effects are unique to each creator and cannot be replicated by AI. Furthermore, in this new system, traditional metrics of success like follower counts may be replaced by new forms of value signaling like "volume transacted" or the "number of times a piece of content is remixed." These crypto-based metrics become a more robust signal of authenticity, interest, and engagement than simple follower numbers.

As creators build their body of work on blockchain systems, they become increasingly antifragile to copying and remixing. Unlike traditional systems where copying might diminish value, in a crypto-based ecosystem, remixing and sharing can actually enhance the original creator's network. A blockchain ledger verifies original authorship, while the act of remixing contributes to network formation around the creator's work. This turns potential threats into opportunities for growth and increased visibility – if an AI model features your work, it becomes a signal of relevance.

In this new paradigm, the creator’s work accounts for only a part of the value they generate. The provenance of their creations, the creator’s cultivated reputation, and the network and transactions surrounding their body of work all contribute significantly to its overall worth and outweigh the value of any standalone piece.

(Image by Jack Butcher)

AI has democratized content creation and allows anyone to become a prolific creator. Yet, this same abundance threatens to erode the perceived value of individual creative works. Crypto-based systems serve as a counterbalance, offering creators a way to establish verifiable, portable reputations and build unique networks around their work. In this new paradigm, an artist's value isn't based solely on their output, but everything else around that output. A creator’s reputation is a store of value, reinforced by the network around said reputation.

As AI becomes a tool for abundance, pushing the boundaries of what can be created, crypto provides the framework for scarcity, attribution, and value capture that doesn’t depend on centralized platforms. The intersection of AI x Crypto ushers in a digital renaissance, empowering creatives to focus on what they do best – create.

We are grateful to the contributors of this post who helped shape our ideas through conversations and who reviewed and provided comments on drafts of this post.

Contributors: Erick Calderon (Snowfro), BoredElonMusk, Ben Ebner, Jon Rogers

Publication: https://paragraph.xyz/@aixcrypto

LJW X/Twitter: https://x.com/WhatTheLJW

Sven X/Twitter: https://x.com/svenwelly

Story Protocol X/Twitter: https://x.com/StoryProtocol

Polychain X/Twitter: https://x.com/polychain

This post is provided for educational and informational purposes only. Nothing written in this post should be taken as financial advice or advice of any kind. The content of this post are the opinions of the authors and not representative of other parties.