A quick overview of how proof-carrying data will reshape the internet

Nexus Chief Scientist, Jens Groth, gave a talk at Devcon 2024 titled “The Verifiability Vision.” 

The talk covers a quick history of zero-knowledge (zk) cryptography, succinct proofs, and then dives into how proof-carrying data can inform the Nexus vision of a Verifiable Internet. 

Or, as Groth says in his talk, “We want to verify the full digital world.”

Check out a full recording of the talk on ETHstream.

There are many applications where zero-knowledge proofs are useful:

• Digital images: With deepfakes on the rise, verifying content like images on the internet is becoming more important. Some cameras can sign the original raw photo. But what you often see in publications is not the original version, but a compressed image. Succinct proofs allow you to verify that the image you see on the internet corresponds to a legitimate compression of an original raw photo with appropriate metadata such as a signature and geotag.
• Online voting: The ability to verify that an encrypted ballot contains a vote for a valid candidate without disclosing which candidate received the vote.
• Verifying an ID or details of an ID: Confirming real-world identifying details while still protecting sensitive data.

The challenge of creating the Verifiable Internet

The technical complexity of implementing zero-knowledge proofs (ZKPs) previously required developers to have deep expertise in cryptography, creating a high barrier to entry. 

However, the advent of zkVMs (zero-knowledge virtual machines) is transforming this landscape.

These zkVMs abstract away much of the cryptographic intricacy, providing developers with familiar environments and tools to build scalable applications that leverage the benefits of ZKPs. 

The prohibitive cost of succinct proof generation was also traditionally another significant obstacle to deploying proofs at scale.

Generating a succinct proof for correctness of a computation is many orders of magnitude more expensive than the computation itself. But ongoing advancements in cryptographic research and hardware optimization have started to bring these costs down dramatically. 

Finally, to create a web of proof-carrying data, we need trustworthy inputs about the real world.

This is called the trust anchor problem. 

While this problem is complex, people are finding creative ways to leverage existing technologies to use trusted data as inputs.

Some examples include, taking identity data from passport chips or repurposing the Transport Layer Security TLS connection in a browser as a witness to the displayed data coming from a trusted organization. 

What does success look like?

Right now most proving demand is coming from blockchain systems looking to scale.

But it also feels like we are at an inflection point and that a Verifiable Internet is becoming plausible. 

“Eventually,” Groth says, “We will have a web of verifiable data. Since verification is cheap, most people want their data to come with proof of correctness. And as the adoption of proof-carrying data continues we will have a much more trustworthy world.”