<100 subscribers
Let’s be honest: You were amazed when you first tried out ChatGPT. That was the point of no return. When OpenAI made its language model public, it didn’t just feel like a new tech demo—it felt like something much bigger. For the first time, machines were speaking with a fluency that felt genuinely human - just like a colleague. But beneath the excitement was a quieter, more unsettling realization: these models weren’t just imitating us. They were absorbing the internet, remixing it, and, soon enough, feeding on their own creations. That was the point, when the shadows on the wall began to change.
Let’s rewind about 2,400 years. In Plato’s famous allegory, a group of people are chained inside a dark cave, facing a wall. Behind them is a fire and between the fire and the prisoners, puppeteers cast shadows on the wall. Since the prisoners can’t turn around, they take the shadows as reality. It’s all they’ve ever known. Then one day, a prisoner breaks free, climbs out into the light, and is overwhelmed by what he sees—the real world and actual objects, color, depth, trees, water, people. He returns to tell the others what he’s discovered. But they don’t believe him. Why would they? To them, the shadows are real and anything else is beyond their imagination.
Now, fast forward to today. We’re still in the cave—but the flickering shadows aren’t cast by firelight anymore. They’re stitched together by algorithms, trained on everything we’ve ever posted, written, filmed or shared. And now, many of those shadows are made by machines imitating older shadows. We're deep in a hall of mirrors—trying to make sense of reflections of reflections. This is where we find ourselves with generative AI today.
The earliest large language models were initially trained on an internet largely filled with human-made content. But that’s changing. Every day, more and more AI-generated content is flooding blog posts, social media, product reviews, scientific literature, even political speeches and probably far more that is out of our imagination. When new models are trained on this output, it creates a strange loop: machines are learning from machines. Researchers call this —“model collapse.” It's like making a photocopy of a photocopy of a photocopy. Eventually, details blur, the signal fades, biases and hallucinations increase and truth and meaning become harder to detect. Not because the models are evil, but because their input is increasingly “synthetic”. The models aren’t learning from the world anymore, they are learning from its own distorted reflection.
Sure, this might sound like a niche concern for data scientists and AI developers. But it’s not. It touches everything—from journalism to science, from education to culture, politics and much more. Everything we are perceiving and doing today (no matter the media type), could even be slightly influenced or completely generated by AI. If most of what we read, watch or interact with online (and probably even offline) is no longer grounded in human experience, what does that mean for us? For how we learn? For how we know what’s real? If high-quality content gets buried under an avalanche of mediocre AI output, how can we protect it?
Enter Web3. At its core, Web3 is about decentralization, trust and verifiability. Not blind trust, but trust with receipts. And that’s exactly what we need when the line between human and machine-made content starts to blur. There are three key advantages that stand out in particular and I would like to line out even further:
Verifying the Subject: Decentralized Identity
Verifying the Object: Content Provenance
Aligning Interests: Preservation of Meaning
Decentralized identity answers questions like: Who made this? Can they be trusted? Are they even human or is it an AI agent behind the content? While content provenance verifies what was made and how (see section below), decentralized identity focuses on who made it and why it should be trusted. Using cryptographic identifiers like DIDs (decentralized identifiers), VCs (verifiable credentials) or newer Web3-native primitives like SBTs (Soulbound-Token), ENS-Names (Ethereum Name Service) or ZKPs (Zero-Knowledge-Proofs), creators can build portable, verifiable reputations and identity across platforms. This isn’t about names—it’s about uniqueness, accountability and integrity.
Importantly, DIDs and VCs can operate without a blockchain. They use cryptographic signatures and secure exchanges of credentials to verify reputation and authenticity. However, when these systems are integrated into a Web3 environment—one built on blockchains—they become more powerful. Blockchains add trustless timestamping, composability across apps and censorship resistance. They make identity proofs portable, persistent and interoperable. While Zero-Knowledge-Proofs enable creators to prove facts without revealing any personal data, Soulbound-Token are non-transferrable digital badges representing skills or achievements that anchor a reputation or skill to a sepecific wallet. ENS-Names are human-readable identifiers tied to wallet addresses which are ending with .eth enabling reputation portability and interoperability across apps. It's also possible to turn existing Web2 domain names into ENS names, which I did with https://h3ss.xyz. With those capabilities you can bridge traditional Web2 identities into the Web3 world, enhancing trust and recognizability. I will also go into this topic in more detail in a later article.
Projects like Worldcoin (developed by Tools for Humanity), BrightID, Proof of Humanity, Gitcoin Passport or Humanity Protocol are systems to verify uniqueness and reputation without requiring centralized gatekeepers. These tools let us verify that a content creator is a real, unique individual with a traceable reputation—even if they remain pseudonymous. The difference is subtle but profound: provenance verifies the object while identity verifies the subject. In an age of deepfakes and bot swarms, knowing that you’re interacting with a real person—not just a well-trained script—becomes crucial. Imagine Reddit, TikTok or YouTube posts tagged "human verified" or X threads signed by pseudonymous authors with rich, portable reputation graphs. Such infrastructure allows for nuance: we can distinguish anonymous experts from spammers and give each the trust they’ve earned and deserve.
Content Provenance focuses on tracking the object such as articles, images, videos or code—clarifying when a piece of media was created, whether it has been altered and how it has evolved over time. Basically the whole “media-lifecycle”. While identity frameworks help determine who made it, provenance frameworks reveal what exactly was made, how it’s changed and whether it can be trusted as an authentic, untampered source. With tools like C2PA, Numbers Protocol, Witness, Rarimo and others creators can attach tamper-proof metadata to their work—proving its origin and integrity. These systems enable decentralized verification of content regardless of where or how it is stored. While some use decentralized storage like IPFS, Filecoin or Arweave for added persistence, the core value lies in the ability to cryptographically prove authorship, timestamping and integrity. Web3 protocols provide the broader framework for such verifiability—ensuring that metadata isn't just present but portable, composable and resistant to manipulation and beyond the reach of centralized gatekeepers. This matters because knowing the provenance helps distinguish original, human-made content from synthetic or manipulated material. It creates a verifiable audit trail—much like tracking components through a supply chain—showing how various media elements were produced, altered and assembled into the final piece. Just as an art collector verifies a painting’s authenticity, humans or AI-agents need to verify the authenticity not only of training data but also of the information, sources and media they consume, process and build upon. This matters more than it sounds. With correctly designed content provenance tools, we can train future AIs on content we trust—content we know came from humans or trusted sources. It’s how we inject sunlight back into the cave.
Why should creators care about provenance and identity? Because Web3 offers more than just verification—it offers motivation and durability through token incentives.
In traditional Web2 platforms, content is monetized indirectly, often through ads and engagement metrics that reward virality over value. In contrast, Web3 introduces mechanisms for direct alignment between creators, curators and consumers. Think creator royalties, community funding and quadratic funding models that prioritize public goods. ENS-linked domain identities, tokenized reputation systems and Soulbound-Tokens can also help communities reward quality contributions. Those contributors like community members of a research DAO or open-source developer could earn digital badges in form of NFTs (transferrable) or SBTs (non-transferrable) as achievements tied to their ENS-name —potentially derived from peer validation or project impact. Those mechanics and tools don’t compensate directly (maybe royalties) but make it possible to reward quality contributions and cultivate trust and continuity. By aligning interests, Web3 helps in such a way to ensure that high-effort, meaningful content—whether produced by humans or advanced AI agents—remains economically viable. Another example: a decentralized publishing platform might reward investigative journalism with collectible tokens or access-based perks, creating an economy around deep, thoughtful content rather than shallow clicks.
Tokens, properly designed smart contracts and DAO structures can support not just content creation but also its curation and preservation. In this context, persistence matters too. Without durable storage and traceable history, valuable content can vanish—especially when platforms change, disappear or being censored. Decentralized networks and protocols provide a path to content permanence and preservation of meaningful content—not just through distributed file storage but through the redundancy and interoperability of open, mainly blockchain-based, ecosystems.
The above mentioned criteria are creating a more resilient and meaningful content ecosystem. It’s not just about fixing the present, it's about ensuring that meaningful digital knowledge can persist and grow not only into the future but also in the right direction.
Besides the examples already mentioned in the sections above, some platforms and projects are already moving in this direction—demonstrating that this isn’t a theoretical future but a living movement. The below mentioned examples are just a small fraction of the vast amounts of projects being build at the moment:
C2PA (backed by Adobe, Microsoft, Intel, BBC, and others) is embedding origin and edit histories into media. TikTok and newsrooms are already testing it to fight fakes and AI-generated misinformation.
Fox Verify, a Layer 2 blockchain project by Fox Corporation, aims to bring cryptographic verification directly into journalism. Articles, images, and sources—stamped and trackable.
Mirror.xyz and Paragraph.xyz allow authors to publish, monetize, and build credibility with embedded Web3 tools like token-gating or minting articles everything linked back to the creators address.
Zora: A protocol for creating, minting and collecting digital art and media with embedded monetization features and provenance, widely used by artists and DAOs in the Ethereum ecosystem.
World: A system that verifies human uniqueness through a biometric proof, anchoring identity to individuals in a privacy-preserving way while enabling global participation in digital economies.
cheqd: A decentralized identity network enabling verifiable credentials with privacy-preserving features and a built-in payment system. It empowers users to control and monetize their identity data, supporting a wide range of use cases—from digital health passports to know-your-customer (KYC) compliance—while ensuring data integrity through cryptographic proofs.
Lens Protocol and Farcaster are decentralized social networks, where you as a user own your content. With these applications your user identity, connections, content and reputation are portable, verifiable and persistent. Within developer- and crypto-native communities they are already gaining significant traction.
ENS: Ethereum Name Service ist the native naming system of the Ethereum Ecosystem and is turning alphanumeric wallet addresses into human-readable names.
Social platforms like LinkedIn, Reddit, Instagram, YouTube, and TikTok are increasingly filled with synthetic media. Content provenance and identity tools could become their new immune system—helping users discern signal from noise.
Gitcoin Grants: A funding platform using quadratic funding to support public goods in open source and climate, already distributing millions in community-directed grants across multiple rounds.
Optimism’s RetroPGF: A mechanism to retroactively reward contributors to the Optimism Collective, reinforcing long-term incentive alignment and value creation in the public goods space.
DNS-linked ENS names: Allow traditional Web2 domain names to be converted into ENS names, offering both continuity and composability in the Web3 ecosystem. This can enhance trust, recognizability and unlock new forms of interoperable, reputation-linked engagement.
These examples show a shared focus: portable identity, verifiable content and sustainable incentives. Not all of these platforms are perfect but they represent tangible progress toward a more trustworthy, identity-centered web. What’s still missing is seamless UX, widespread adoption and tighter integration into existing products. These tools need to be invisible in the best way: working quietly in the background, as obvious as a timestamp.
Plato’s cave ends with a moment of painful clarity—the realization that the shadows we once trusted were illusions. And today? We're facing a similar awakening. The generative AI era isn’t just a marvel of progress; it's a collapse of meaning, a hyperreal echo chamber that trades truth and trust for speed and abundance. And we’re maybe sleepwalking through it—fascinated by the shadows that talk back.
Web3 cracks the wall - it offers more than defense: design, direction and durability. A new digital architecture where provenance is the default, identity is programmable and incentives are designed to reward transparency, accountability and meaningful contribution—whether the source is human or AI.
This isn’t about rejecting AI or pining for the past. It’s about building a future where we don’t just consume content—we understand it. Where knowing who said what, when and why isn’t a luxury; it’s a baseline. Where verification isn’t a blue check you buy from X but a protocol-native public good. The next internet will not be won with spectacle and a singular moment everybody will remember. It will be won with structure and rock-solid foundations. We don’t need to burn the cave down. We just need to stop mistaking shadows for light.
The new internet is already here. Let’s make it work.
