Why Web3 and the AI-Internet Belong Together
A tweet this week from Chris Frantz, the founder of email platform Loops, irked me.
Frantz said “90% of the people I know in web3 have pivoted their company to AI.”
What got to me was not that founders are so obsessed with securing venture capital that they’ll glob onto the next “in” thing. (Let’s save that problem of Silicon Valley fickleness for another time.) It was that people see the various elements of the complex new digital economy forming around us – artificial intelligence, blockchain, the metaverse, programmable money, digital identity, cryptographic proofs, quantum computing, IoT and so forth – as unrelated, exchangeable pieces, when they’re really intertwined and complementary.
You’re reading Money Reimagined, a weekly look at the technological, economic and social events and trends that are redefining our relationship with money and transforming the global financial system. Subscribe to get the full newsletter here.
AI will need Web3, and vice versa. Why not build both?
I see this harmful, reductionist simplification of “Web3,” “blockchain” and “crypto” stemming from a fundamental failure to identify the core, unifying feature of all projects that end up under those labels. To me, the commonality is that they all use a novel system of distributed record-keeping and incentives to deal with the basic problem of human trust around information. These technologies help communities of mistrusting strangers collectively maintain open data records that allow them to distribute and share valuable (or sensitive) information among themselves without the intermediation of middlemen.
In addressing how to keep valuable information secure in a decentralized environment, Web3, crypto and blockchain address a societal challenge that has been with us since the start of the internet. But, now, in the age of AI, when uncertainty over information is going to go stratospheric, it’s an even more urgent matter.
So why do people – whether they’re founders flipping from one fad to the next, or policymakers who think crypto is just for money-laundering – not comprehend the all-encompassing importance of this new data architecture?
At the risk of being sacreligious, I think it goes back to cryptocurrency's roots, with the founding of Bitcoin.
Back then, the messaging should have been about information, about sharing data, about protecting privacy – in essence, the primary concerns of the cypherpunks whose mailing list Satoshi Nakamoto used to unveil the Bitcoin white paper in October 2008.
Now, I’m not really blaming Satoshi here. The founder was simply offering up one of a number of information solutions that the cypherpunks had ruminated on for years: a cryptography-based digital currency. Satoshi knew that, while money is special in terms of its fundamental importance to society, it is really just another form of information.
Money isn’t a thing. It’s a standardized, commonly agreed-upon symbolic representation of value. It’s a particular type of information that, because it is highly valued, requires an elaborate, institutionalized system to instill confidence that people and entities won’t abuse it. But it is far from the only type of information that carries value and which, by extension, requires institutional coordination. That’s why Bitcoin was, to me, a prototype for a much bigger idea.
There were, of course, many early Bitcoin believers, including Ethereum founder Vitalik Buterin, who got it. They recognized that this decentralized data architecture could be applied to the myriad problems we face in sharing valuable information in the digital age.
The problem was that in the general public’s eye, as well as those of regulators who wished to jam this strange square peg into the round hole of traditional finance, cryptocurrencies and blockchains were purely about money.
That misunderstanding has set us back, perpetuating a Web2 structure of harmful data manipulation by giant internet platforms that have sown mistrust in our information systems and democracy. If a broader understanding of the potential had existed, this industry would have addressed its inherent scaling, legal and privacy challenges more readily. Maybe there’d have been less of an instinct toward scams and “number go up” token casinos, and more drive to build meaningful solutions to the world’s problems.
But, now? Now, in the age of artificial intelligence, this misunderstanding gets downright dangerous.
Please don’t accuse me of naive “blockchain fixes this” hand-waving. The challenges of the AI moment are daunting, from protecting copyright in the inputs of large language models (LLMs), to avoiding racial bias in their outputs, to the “liar’s dividend” fostered by our current inability to distinguish between real content and AI-created fabrications. There is no easy way to save humanity from the machines. Whatever solution arises will inevitably draw on a wide range of technologies and policies.
Here’s what I do know: we are not going to resolve these matters with an outdated 20th-century regulation-technology stack. We need a decentralized governance system for how we produce, verify, and share information in this new era.
How it can help
Whether or not, as currently designed, they can deliver what’s needed, blockchains do have qualities that can help.
Immutable ledgers allow us to track the provenance of images and other content and so could protect against deep fakes. The same might apply to testing the integrity of the datasets on which machine learning AI products are trained. Cryptocurrencies could be used to pay people worldwide, in a borderless digital manner, for their contributions to AI training. Projects such Bittensor are building tokenized, blockchain-government communities that incentivize AI developers to build human-friendly models (addressing the concern that AI systems owned by private corporations are incentivized to put the profits of shareholders over the rights of users.)
There’s a long way to go before these ideas can deliver on this promise at the scale that’s needed, if they ever will. Also, success will require the integration of a range of other technologies – zero-knowledge proofs, homomorphic encryption, secure computing, digital identities and decentralized credentials, IoT – as well as smart, multi-stakeholder legislation that protects privacy, punishes bad behavior and encourages human-centric innovation.
But to position Web3, blockchain, crypto, or whatever you want to call it, as a has-been with no place in the emerging digital future is to severely mis-comprehend the problem at hand.