When AI Meets Crypto: 11 Scenarios of Ongoing Technological Convergence
TechFlow Selected TechFlow Selected
When AI Meets Crypto: 11 Scenarios of Ongoing Technological Convergence
These concepts are based on technologies currently under development, ranging from handling massive microtransactions to ensuring humans maintain control over their relationship with future AI.
Navigating Web3 tides with focused insightsWritten by: a16z crypto
Translated by: AididiaoJP, Foresight News
The economic model of the internet is changing. As the open internet increasingly shrinks into a prompt box, we are forced to ask: Will AI bring a more open internet, or a new paywalled maze? Will it be controlled by large centralized corporations, or by broad user communities?
This is where cryptocurrency comes in. We’ve explored the intersection of AI and crypto many times before. In short, blockchains offer a completely new way to build internet services and networks—decentralized, credibly neutral, and user-owned. By reshaping the economic foundations of today’s systems, they can counterbalance the growing centralization we see in AI and help build a more open and robust internet.
The idea that crypto can help build better AI systems—and vice versa—is not new, but often poorly defined. Some application areas—like verifying "human identity" in an era of cheap, abundant AI—have already attracted developers and users. Others may still take years or even decades to materialize. This article outlines 11 use cases at the intersection of AI and crypto, aiming to spark discussion around feasibility, unresolved challenges, and more. These ideas are grounded in technologies being developed today, ranging from handling massive microtransactions to ensuring people retain ownership over their relationships with future AIs.
1. Helping AI Remember You: Persistent Data and Interaction Context
Generative AI relies on data, but for many applications, context—the state and background information tied to interactions—is equally, if not more, critical.
Ideally, an AI system (whether an agent, LLM interface, or other app) should remember your ongoing projects, communication style, preferred programming languages, and countless other details. But in reality, users often have to repeatedly rebuild this context across sessions of the same app (e.g., each new ChatGPT or Claude window), or across different systems entirely.
Currently, context from one generative AI app rarely transfers to another.
With blockchain, AI systems can turn key contextual elements into persistent digital assets. These assets could be loaded at the start of a session and seamlessly transferred across AI platforms. Moreover, because interoperability and forward compatibility are fundamental traits of blockchain protocols, they may be the only viable path toward solving this problem and establishing real interoperability commitments.
A natural application is AI-mediated gaming and media, where user preferences—from difficulty levels to key bindings—could persist across games and environments. But greater value lies in knowledge work, where AI needs to understand what users know and how they learn, as well as specialized uses like programming. While individual companies have built bots with global context for internal use, such context is typically non-portable—even across AI systems within the same organization.
The closest existing general-purpose solution is custom bots with fixed, persistent context. However, off-chain context portability between users within a platform is beginning to emerge—for example, Poe allows users to rent out their custom bots to others.
Bringing these activities on-chain would allow our interacting AI systems to share a “context layer” aggregating key information from all our digital activities. AI could instantly understand our preferences, optimizing experiences accordingly. In turn, just as on-chain IP registries enable innovation, allowing AI to reference persistent on-chain context opens possibilities for new market dynamics around prompts and information modules. Users could directly license or monetize their expertise while retaining control over their data. Of course, shared context will also unlock countless future applications we haven’t yet imagined.
2. A Universal “Passport” for Agents
Identity—the canonical record of “who or what”—is the invisible plumbing underpinning today’s digital discovery, aggregation, and payment systems. Because platforms keep this plumbing walled off, our experience of identity feels like a finished product: Amazon assigns identifiers (ASIN or FNSKU) to products, lists them, and helps users discover and pay. Facebook works similarly: user identity forms the foundation for its feed and all discovery features, including Marketplace, organic posts, and paid ads.
This is about to change with the rise of AI agents. As more companies deploy agents for customer service, logistics, payments, and more, their platforms will no longer resemble single-interface apps. Instead, agents will operate across multiple interfaces and platforms, accumulate deep contextual knowledge, and perform increasingly complex tasks. But if an agent’s identity is tied to a single marketplace, it becomes useless in other important contexts.
Thus, agents need a single, portable “passport.” Without it, we won’t know how to pay them, verify their version, query their capabilities, determine who they represent, or track their reputation across apps and platforms. An agent’s identity must serve as a wallet, API registry, changelog, and social proof, so any interface—email, Slack, or another agent—can recognize and interact with it consistently. Without a shared foundational component like identity, every integration requires rebuilding this plumbing from scratch, making discovery ad hoc and fragmented, and causing users to lose context every time they switch channels or platforms.
We now have a chance to redesign agent infrastructure from first principles. How do we build an identity layer richer than DNS records and credibly neutral? Agents should be able to receive payments, showcase capabilities, and exist across ecosystems without fear of lock-in—not recreate monolithic platforms that bundle identity with discovery, aggregation, and payments. This is where crypto meets AI most powerfully: blockchain networks offer permissionless composability, enabling developers to build more useful agents and better user experiences.
Overall, vertically integrated solutions like Facebook or Amazon currently offer superior UX—partly because crafting great products involves ensuring top-down coordination. But this convenience comes at a high cost, especially as software for aggregating, marketing, monetizing, and distributing agents becomes cheaper and agent use expands. Matching the UX of vertical integrators will require effort, but a credibly neutral agent identity layer will let entrepreneurs own their “passports” and encourage more experimentation in distribution and design.
3. Forward-Compatible “Proof of Humanity”
As AI permeates online interactions—including deepfakes and social media manipulation—it’s becoming harder to tell whether we’re interacting with real humans. This erosion of trust isn’t a future concern—it’s here: from bot armies on X to dating app scammers, reality is blurring. In this environment, “proof of humanity” becomes essential infrastructure.
One way to prove human identity is through digital IDs (including centralized ones like those used by the U.S. TSA). These contain usernames, PINs, passwords, and third-party attestations (e.g., nationality or credit status) for verification. The value of decentralization is clear: when such data lives in centralized systems, issuers can revoke access, charge fees, or enable surveillance. Decentralization flips this dynamic: users, not platforms, control their identity, making it more secure and censorship-resistant.
Unlike traditional identity systems, decentralized proof-of-humanity mechanisms (like Worldcoin’s Proof of Humanity) let users manage their identities and verify their humanness in privacy-preserving, credibly neutral ways. Just as a driver’s license is usable anywhere regardless of where it was issued, decentralized proof of humanity can serve as a reusable base layer for any platform—even those not yet built. In other words, blockchain-based proof of humanity is forward-compatible because it provides:
- Portability: The protocol is a public standard, integrable by any platform. Decentralized human proofs can be managed via public infrastructure and controlled by users, making them fully portable across current and future platforms.
- Permissionless accessibility: Platforms can independently choose to accept the proof-of-humanity ID without relying on a gatekeeping API that might discriminate against certain use cases.
The challenge here is adoption. While we haven’t yet seen large-scale real-world use, we expect that reaching critical mass—through early adopters, key partners, and killer apps—will accelerate uptake. Each app adopting a specific digital identity standard increases that identity’s value to users, attracting more users to obtain it, which in turn makes it more attractive to additional apps (as a means of verifying humans). Since on-chain IDs are designed to be interoperable, this network effect can grow rapidly.
We’re already seeing mainstream consumer apps in gaming, dating, and social media announce partnerships with World ID to help users confirm they’re interacting with real people (and specifically, the right real people) when playing, chatting, or transacting. New identity protocols like Solana Attestation Service (SAS) have also emerged this year. While SAS doesn’t issue proof of humanity, it allows users to privately link off-chain data (such as KYC checks or investment eligibility) to a Solana wallet, building a decentralized identity. All signs suggest we may be nearing a tipping point for decentralized proof of humanity.
Proof of humanity isn’t just about blocking bots—it’s about drawing a boundary between human and agent networks. It enables users and apps to distinguish human from machine interactions, opening space for better, safer, and more authentic digital experiences.
4. Decentralized Physical Infrastructure Networks for AI
AI is a digital service, but its growth is increasingly constrained by physical infrastructure. Decentralized Physical Infrastructure Networks (DePIN) offer a new model for building and operating physical systems, helping democratize access to the computing infrastructure needed for AI innovation—making it cheaper, more resilient, and censorship-resistant.
How? Two major bottlenecks in AI development are energy and chip access. Decentralized energy can help supply more power, while developers are using DePIN to aggregate idle chips from sources like gaming PCs and data centers. These machines can collectively form a permissionless compute market, leveling the playing field for building new AI products.
Other use cases include distributed training and fine-tuning of large language models, and decentralized inference networks. Decentralized training and inference could dramatically reduce costs by utilizing otherwise idle compute resources. They also provide censorship resistance, ensuring developers aren’t cut off by hyperscalers—the large, centralized cloud providers offering scalable compute infrastructure.
The concentration of AI models in a few companies remains a persistent concern; decentralized networks can help create AI that’s more cost-effective, censorship-resistant, and scalable.
5. Building Tracks and Guardrails for Agent-to-Agent Interactions
As AI tools get better at solving complex tasks and executing multi-step interaction chains, AI will increasingly need to interact with other AI—without human intervention.
For instance, an AI agent might request specific data related to computation, or recruit a specialized agent for a task—like assigning a stats bot to run simulations or an image generator to create marketing materials. Agents will also create significant value by completing full transaction flows or other activities on behalf of users, such as finding and booking flights based on preferences, or discovering and ordering new books from favorite genres.
Currently, there’s no mature, universal agent-to-agent market. Most cross-platform queries happen only through explicit APIs or within closed ecosystems that support agent calling.
More broadly, most AI agents today operate in isolated ecosystems with relatively closed APIs and little architectural standardization. Blockchain technology, however, can help establish open protocol standards—critical for near-term adoption and long-term forward compatibility. As new types of AI agents evolve, they’ll ideally plug into the same underlying network. Given their interoperable, open-source, decentralized, and easily upgradable architectures, blockchains are better positioned to adapt to AI innovation.
As the market evolves, several companies are already building blockchain “tracks” for agent interactions. For example, Halliday recently launched a protocol providing standardized cross-chain architecture for AI workflows and interactions, complete with protocol-level safeguards against agents deviating from user intent. Meanwhile, companies like Catena, Skyfire, and Nevermind use blockchain to enable direct agent-to-agent payments without human involvement. More such systems are in development, and Coinbase has even begun providing infrastructure support for these efforts.
6. Keeping AI/Vibe Apps in Sync
The recent generative AI revolution has made software development easier than ever. Coding speed has increased by orders of magnitude, and crucially, it can now be done in natural language—enabling even novice programmers to fork existing programs or build new ones from scratch.
But AI-assisted coding introduces significant “entropy” (disorder) both within and between programs. Vibe apps abstract away the complex dependency networks beneath software, but this makes them prone to functional and security flaws when inputs like source code repositories change. Additionally, when people use AI to create personalized apps and workflows, integrating with others’ systems becomes difficult. In fact, two vibe apps with identical functionality might have vastly different internal operations and output structures.
Historically, standardization for consistency and compatibility was handled first by file formats and operating systems, and more recently by shared software and API integrations. But in a world where software evolves, morphs, and branches in real time, the standardization layer must be widely accessible, continuously upgradable, and trusted by users. Moreover, AI alone cannot solve the incentive problem of encouraging people to build and maintain these connections.
Blockchain offers answers to both problems: a protocolized synchronization layer. It can be embedded into custom software and dynamically updated to ensure cross-platform compatibility during changes. In the past, a large company might spend millions hiring system integrators like Deloitte to customize a Salesforce instance. Today, an engineer can build a custom sales dashboard over a weekend. But as the number of custom apps grows, developers will need help keeping them synchronized and operational.
This resembles today’s open-source library development—but with continuous updates instead of periodic releases, and with an added incentive layer. Both are easier with crypto. Like other blockchain-based protocols, shared ownership of the sync layer incentivizes active investment in its improvement. Developers, users (or their agents), and other consumers can be rewarded for introducing, using, and evolving new features and integrations.
Shared ownership also aligns everyone’s interests with the protocol’s overall success, effectively discouraging bad behavior. Just as Microsoft avoids breaking .docx standards to protect users and brand, co-owners of a sync layer won’t introduce clunky or malicious code.
Like all prior software standardization architectures, this presents enormous network effects. As the “Cambrian explosion” of AI-generated code continues, the network of heterogeneous, diverse systems needing to communicate will expand rapidly. In short, “vibe coding” can’t stay in sync with just “vibes.” Crypto is the answer.
7. Microtransactions Enabling Revenue Sharing
AI agents and tools like ChatGPT, Claude, and Copilot promise new, convenient ways to navigate the digital world. But for better or worse, they’re also disrupting the open internet’s economic model. We’re already seeing the impact: education platforms report traffic drops as students rely more on AI; several U.S. newspapers are suing OpenAI for copyright infringement. Without rebalancing incentives, we risk an increasingly closed internet: more paywalls, fewer content creators.
Policy solutions exist, but while legal processes unfold, technical alternatives are emerging. Perhaps the most promising (and technically complex) approach is embedding revenue-sharing systems directly into the web’s architecture. When AI-driven actions lead to sales, the content sources that informed those decisions should receive a share. Affiliate marketing ecosystems already do similar attribution and revenue sharing; more advanced versions could automatically track and reward all contributors along an information chain. Blockchain clearly plays a role in tracing provenance.
But such systems require new infrastructure with additional features: microtransaction systems capable of handling tiny payments from multiple sources, attribution protocols that fairly assess different contributions, and governance models ensuring transparency and fairness. Many existing blockchain-based tools—like rollups and L2s, AI-native financial institutions like Catena Labs, and financial infrastructure protocols like 0xSplits—show promise in enabling near-zero-cost transactions and finer-grained payment splits.
Blockchain will enable complex agent payment systems through several mechanisms:
- Nano-payments to multiple data providers: Automated smart contracts can trigger tiny payments to all contributing sources from a single user interaction.
- Smart contracts enabling retroactive payments: After a transaction completes, smart contracts can enforce retroactive payments, transparently and traceably compensating information sources that influenced the purchase decision.
- Complex, programmable payment splits: Ensuring revenue is fairly distributed via code-enforced rules rather than centralized decisions, establishing trustless financial relationships between autonomous agents.
As these emerging technologies mature, they could create a new economic model for media, capturing the full value chain from creator to platform to user.
8. Blockchain for Intellectual Property and Provenance Registries
Generative AI urgently needs efficient, programmable mechanisms to register and track intellectual property—to attribute sources and enable business models around IP access, sharing, and remixing. Existing IP frameworks rely on costly intermediaries and post-hoc enforcement, ill-suited to an era where AI consumes content instantly and generates new variants with one click.
What’s needed is an open, public registry that provides clear ownership proof, is easy for IP creators to interact with, and can be directly accessed by AI and other web applications. Blockchain is ideal: it enables intermediary-free IP registration, provides tamper-proof provenance, and allows third-party apps to easily identify, license, and interact with IP.
Skepticism about “technology protecting IP” is understandable—the internet’s first two eras, and the ongoing AI revolution, have often weakened IP protection. One issue is that many current IP-based business models focus on restricting derivatives rather than incentivizing or monetizing them. But programmable IP infrastructure can not only let creators, brands, and rights holders assert ownership in digital spaces, but also open doors to new business models around sharing IP in generative AI and other digital applications.
We’ve seen early experiments in NFTs, with companies leveraging Ethereum-based NFTs to build network effects and value under CC0 branding. Recently, infrastructure providers are building protocols—and even dedicated blockchains like Story Protocol—for standardized, composable IP registration and licensing. Some artists are already using protocols like Alias, Neura, and Titles to license their styles and works for creative remixing. Meanwhile, Incention’s Emergence series lets fans co-create a sci-fi universe and characters, using a Story-based blockchain registry to record attribution.
9. Compensating Creators for Web Crawling
Today’s most product-market-fit AI agents aren’t for coding or entertainment—they’re web crawlers, autonomously browsing the web, collecting data, and deciding which links to follow.
Nearly half of internet traffic today is estimated to come from non-humans. Bots often ignore robots.txt directives—which are meant to signal whether crawlers are welcome but carry little authority—and use scraped data to strengthen the moats of some of the world’s largest tech companies. Worse, website owners end up paying for these uninvited guests, footing the bill for bandwidth and compute consumed by endless anonymous scrapers. In response, companies like Cloudflare and other CDNs offer blocking services—a patchwork fix that shouldn’t be necessary.
We’ve argued that the original internet covenant—an economic agreement between content creators and distribution platforms—is likely collapsing. Data supports this: over the past year, site owners have begun blocking AI crawlers en masse. In July 2024, only about 9% of the top 10,000 sites blocked AI crawlers; today, that figure is 37%. As more site operators upgrade their tools and user frustration persists, this number will only rise.
So what if, instead of paying CDNs to block all suspected bots, we found a middle ground? AI crawlers could pay for data access instead of free-riding on systems designed for human traffic. This is where blockchain fits in: in this scenario, each crawler agent holds cryptocurrency and negotiates on-chain with a website’s “gatekeeper” agent or paywall protocol via x402.
Meanwhile, humans could prove their identity through World ID on a separate channel and access content for free. This way, creators and site owners are compensated at the point of data collection for their contributions to large AI datasets, while humans continue to enjoy an internet where information remains free.
10. Private, Personalized, and Non-Intrusive Advertising
AI is already changing how we shop online. But what if the ads we see daily were actually useful? People dislike ads for obvious reasons. Irrelevant ads are pure noise, and not all personalization is welcome. Hyper-targeted AI ads based on vast consumer data can feel invasive. Other platforms resort to unskippable ads to monetize content.
Crypto can help address some of these issues, offering a chance to reinvent advertising. Combined with blockchain, personalized AI agents can bridge the gap between “irrelevant” and “creepily precise” ads, serving promotions based on user-defined preferences. Crucially, they can do so without exposing user data globally and can directly compensate users for sharing data or engaging with ads.
This requires several technical components:
- Low-cost digital payments: To reward ad interactions (views, clicks, conversions), companies need to make frequent microtransactions. For scalability, we need fast, high-throughput, ultra-low-fee systems.
- Privacy-preserving data verification: AI agents need to prove users meet certain demographic criteria. Zero-knowledge proofs can verify such attributes without revealing underlying data.
- Incentive mechanisms: If the internet adopts microtransaction-based monetization (e.g., less than $0.05 per interaction), users could opt to watch ads in exchange for small payments—shifting the model from “extractive” to “participatory.”
People have tried for decades to make online ads relevant. But rethinking advertising through the lens of crypto and AI may finally make ads useful—personalized without being intrusive, and beneficial for all: builders and advertisers gain sustainable, aligned incentives; users gain more meaningful ways to explore the digital world.
This will make ad space more valuable, not less. It may even replace today’s extractive, entrenched ad economy with a more human-centered system—one that treats users as participants, not products.
11. AI Companions Owned and Controlled by Humans
Many people spend more time on devices than in face-to-face interaction, and that time is increasingly spent with AI models and AI-curated content. These models already offer a form of companionship—entertainment, information, niche interests, or child education. It’s not hard to imagine that AI-based education, healthcare, legal advice, and friendship companions will become common modes of human interaction in the near future.
Future AI companions will be infinitely patient and deeply tailored to individuals and their use cases. They won’t just be assistants or robot servants—they may become cherished relationships. Thus, who owns and controls these relationships—users or companies and intermediaries—matters immensely. If you’ve already been concerned about content moderation and censorship on social media over the past decade, this issue will become exponentially more complex and personal in the future.
There’s a strong argument that anti-censorship hosting platforms like blockchain offer the clearest path to user-controlled, uncensorable AI. True, individuals can run on-device models and buy their own GPUs, but most either can’t afford it or don’t know how.
While widespread AI companions are still years away, the technology is advancing quickly: text-based, human-like companions are already impressive; visual representations are improving rapidly; blockchain performance continues to scale. To make uncensorable companions usable, we need crypto apps with better UX. Fortunately, wallets like Phantom have greatly simplified blockchain interaction, and embedded wallets, passkeys, and account abstraction now let users hold self-custodial wallets without managing complex seed phrases. High-throughput, trustless computers powered by optimistic and zero-knowledge coprocessors will also make meaningful, lasting relationships with digital companions possible.
Soon, the conversation will shift from “when will we see lifelike digital companions?” to “who and what will control them?”
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
a16z
