Real AI & crypto use cases, No. 4 – Cointelegraph Magazine

Each day this week we’re featuring a real-life use case for artificial intelligence and cryptocurrency—including reasons why you don’t necessarily believe the hype. Today: How blockchain can fight fakes.

Generative AI is very good at producing fake photos, fake letters, fake bills, fake conversations—fake everything. Co-founder Illia Polosukhin warns that soon we won’t know what to trust.

“If we don’t solve reputation and content authentication (issues), things are going to get really weird,” Polosukhin explained. “You’ll get a call and you’ll think it’s from someone you know, but it’s not.”

“Every image, every content, book you see is going to be (suspicious). Imagine children studying a history book, when in fact each child has seen a different textbook – it is trying to teach a specific way to influence them.”

Blockchain can be used to transparently trace the origin of online content so that users can differentiate between real content and AI-generated images. But it cannot tell the difference between truth and lies.

“That’s the wrong view of the problem because people write things that are untrue all the time. More importantly, when you see something, was it seen by the person it was describing?” Polosukhin said.

“This is where reputation systems come in: OK, this content comes from this author; can we trust what the author says?”

“So cryptography becomes a tool to ensure consistency and traceability, and then you need the reputation around that cryptography – the on-chain accounts and record keeping to really ensure that ‘X published this’ and ‘X is now working for Cointelegraph job’.”

If this is a good idea, why doesn’t anyone do it?

There are a variety of existing supply chain projects that use blockchain to prove the provenance of goods in the real world, including VeChain and OriginTrail.

However, content-based provenance has yet to take off. The Trive News project, which aimed to crowdsource article verification via the blockchain, and the Po.et project, which marked a transparent history of content on the blockchain, both no longer exist.

recent, de facto agreement Launched, combining artificial intelligence and Web3 technology in an attempt to crowdsource news verification.The program joined the Content Authenticity Initiative in March last year

When someone shares an article or piece of content online, it is first automatically verified using artificial intelligence, then double-checked by fact-checkers in the protocol, and then recorded on-chain along with the timestamp and transaction hash. .

“We will not republish content on our platform, but we will create a permanent on-chain record of it, along with a record of the fact checks and validators that were performed,” founder Mohith Agadi told The Decrypting story.

In August, global news agency Reuters ran a proof-of-concept pilot program that used a prototype Canon camera to store metadata for photos on the chain in compliance with the C2PA standard.

It also integrates Starling Lab’s authentication framework into its image desktop workflow. Through the metadata, edit history, and blockchain registration embedded in the photo, users can verify the authenticity of the photo by comparing their unique identifier with the identifier recorded on the public ledger.

Academic Research Work is also ongoing in the area.

Is blockchain needed?

Technically, no. One of the issues holding back this use case is that you don’t actually need blockchain or cryptography to prove the origin of a piece of content. However, doing so makes the process more robust.

So while you can use cryptographic signatures to verify content, Polosukhin asks readers how to be sure this is the correct signature? If the key is posted on the original website, someone can still hack the website.

Web2 handles these issues by using trusted service providers, he explained, “but there are always problems.”

“Symantec was hacked and they issued invalid SSL certificates. Sites are being hacked – Curve, even Web3 sites are being hacked because they’re running on a Web2 stack,” he explain.

“So, at least from my perspective, if we expect it to be used in a malicious way in the future, we need tools that can really adapt to that.”

Don’t believe the hype

People have been discussing the use cases of blockchain to combat “disinformation” and deepfakes long before artificial intelligence took off, but only recently have progress been made.

Microsoft has just launched a new watermark to combat generated artificial intelligence fraud used in election campaigns. A watermark from the Alliance for Authenticity of Content Origins is permanently attached to the metadata and shows who created it and whether artificial intelligence was involved.

The New York Times, Adobe, BBC, Truepic, Washington Post and Arm are members of C2PA. However, this solution does not require the use of a blockchain, as metadata can be protected using hashes and certified digital signatures.

That said, it can also be recorded on the blockchain, as Reuters’ pilot program in August demonstrated. C2PA’s awareness arm is called the Content Authenticity Initiative, while Web3 agencies, including Rarible, Fact Protocol, Livepeer and Dfinity, He is a member of CAI flying the blockchain flag.

Also read:

Real Use Cases of Artificial Intelligence in Cryptocurrency, First: The Best Currency for Artificial Intelligence is Cryptocurrency
Real Artificial Intelligence Use Cases in Crypto, Part Two: Artificial Intelligence Can Run DAOs
Real-life AI use cases in crypto, No. 3: Smart contract auditing and network security

subscription

The most fascinating read in blockchain. Delivery is weekly.

Andrew Fenton

Andrew Fenton lives in Melbourne and is a journalist and editor covering cryptocurrency and blockchain. He has been national entertainment writer for News Corp Australia, film reporter for SA Weekend and reporter for Melbourne Weekly.

Follow the author @AndrewFenton