Generative AI excels at creating fake photos, letters, bills, and conversations, posing a significant challenge to discerning genuine content. Near co-founder, Illia Polosukhin warns that without solutions for content reputation and authentication, the digital landscape will become increasingly deceptive. Blockchain technology offers a potential solution by tracing the provenance of online content, enabling users to distinguish between authentic and AI-generated material.
Blockchain and AI: Battling the Surge of Fake Content
“If we don’t solve this reputation and authentication of content (problem), shit will get really weird,” Polosukhin explains. “You’ll get phone calls, and you’ll think this is from somebody you know, but it’s not.”
“All the images you see, all the content, the books will be (suspect). Imagine a history book that kids are studying, and literally every kid has seen a different textbook — and it’s trying to affect them in a specific way.”
Blockchain can be used to transparently trace the provenance of online content so that users can distinguish between genuine content and AI-generated images. But it won’t sort out truth from lies.
“That’s the wrong take on the problem because people write not-true stuff all the time. It’s more a question of when you see something, is it by the person that it says it is?” Polosukhin says.
“And that’s where reputation systems come in: OK, this content comes from that author; can we trust what that author says?”
“So, cryptography becomes an instrument to ensure consistency and traceability and then you need reputation around this cryptography — on-chain accounts and record keeping to actually ensure that ‘X posted this’ and ‘X is working for Cointelegraph right now.’”
Seriously, don’t trust anything you see online anymore. Faking stuff is trivial. You cannot tell the difference. There are no watermarks, and watermarks can be defeated easily. This genie is not going back in the bottle.
— Ethan Mollick (@emollick) November 27, 2023
If it’s Such a Great Idea Why isn’t Anyone Doing it Already?
Blockchain projects like VeChain and OriginTrail prove the provenance of physical goods, but content-based provenance remains underdeveloped. Trive News and Po.et attempted blockchain-based article verification but are now defunct.
Recently, Fact Protocol emerged, combining AI and Web3 to crowdsource news validation. They joined the Content Authenticity Initiative in March 2023.
“When someone shares an article, it’s first validated by AI, then fact-checkers verify it. The information, along with timestamps and transaction hashes, is recorded on-chain,” says founder Mohith Agadi.
In August, Reuters piloted a program using a prototype Canon camera to store photo metadata on-chain using the C2PA standard. They integrated Starling Lab’s authentication framework into their workflow, allowing users to verify a picture’s authenticity through its unique identifier on the public ledger.
Why did @FactProtocol choose blockchain technology to build a decentralized fact-checking system? A thread 🧵👇
1. Tamper-Proof Records:
Blockchain’s immutability ensures that fact-checking records on Fact Protocol cannot be altered or tampered with, providing a secure and… pic.twitter.com/tEOrl4Us4o
— Fact Protocol (@FactProtocol) November 21, 2023
Academic Research and Technological Needs
While blockchain provides a robust solution, it’s not the only way to verify content origins. Cryptographic signatures alone can serve this purpose, though they are vulnerable to hacking if the key is posted on the originating website.
Web2 addresses these issues using trusted service providers, but they are also susceptible to breaches. For example, Symantec’s SSL certificates were compromised, and websites, including Web3 sites running on Web2 infrastructure, are frequently hacked.
“We need resilient tools to prevent future malicious uses,” Polosukhin emphasizes.
Current Blockchain Developments and Future Directions
Discussions about blockchain’s role in combating disinformation and deep fakes predate AI’s rise, but progress has been slow. Microsoft recently introduced a watermark to prevent generative AI fakes in election campaigns. The watermark, from the Coalition for Content Provenance Authenticity (C2PA), attaches permanently to metadata, indicating the creator and whether AI was involved.
C2PA members include The New York Times, Adobe, BBC, Truepic, Washington Post, and Arm. While blockchain isn’t necessary for this solution, metadata can be secured with hashcodes and digital signatures. Reuter’s August pilot showed blockchain’s potential, and the Content Authenticity Initiative, part of C2PA, includes Web3 entities like Rarible, Fact Protocol, Livepeer, and Dfinity promoting blockchain use.
In conclusion, while blockchain can significantly enhance content authenticity and traceability, the real challenge lies in verifying the truthfulness of the content itself. With ongoing advancements and collaborations, the future may hold more robust solutions to combat the growing threat of fake content.
Leave a Reply