arweave-world.png
2023-08-17
By Adeola
Tip the author

Arweave Protocol Working To Solve One Of AI’s Biggest Challenges

There are indications that the challenge that is bedeviling generative artificial intelligence across the world is being worked on by Arweave and is expected to be launched soon, that is according to Sam Williams, founder of Arweave and CEO of Forward Research.

Generative artificial intelligence, a use case of artificial intelligence which is capable of generating text, images and media by using generative models, has gained popularity since the year 2020 till date with investment by big tech companies including Microsoft, Google and Baidu investing companies that develop generative models.

But the potentials of this innovation have also raised issues of cybercrime, fake news and deep fakes. The difficulty in identifying AI generated text, image or media has resulted in many being deceived or manipulated.

Early this year, pictures of Pope Francis in a puffer jacket went viral on the internet and fooled many people. The creator of the image, Pablo Xavier (31), said he was trying to be funny. Another image of the former President of the United States fighting off the New York Police also went viral.

However, Arweave is coming up with a solution to this problem, said Sam Williams, founder of Arweave and CEO of Forward Research. Sam revealed this at the Arweave Day - Warsaw discussion held on X space. Arweave Day - Warsaw is an event planned for the Eth Warsaw community during the Warsaw Blockchain week. Arweave is one of the sponsors of the Eth Warsaw event and hackathon and Sam Williams as well as many from the Arweave community are expected to attend.

The solution, called Responsible AI Ledger is a system that allows a hash to be attached to generative AI outputs.

“If you are an AI or a synthetic data platform, when you output a piece of data to the user, you can upload a hash of that piece of that data or a fingerprint of some form to Arweave. And when you do so, you are essentially de-risking one of the main problems of AI safety which is what if someone creates fake news or deep fake photographs of political events and you can’t tell what is real and what is fake on the internet?” Williams asked.

Williams said that the problem could be solved if large AI platforms store publicly on Arweave a hash of contents generated with AI in order to allow anyone to look up the hash on Arweave using query engines.

He said discussions were on with platforms about the idea and it could take some time before production starts.

The founder, revealing other plans in Arweave’s pipeline said a user interface for recently announced universal content marketplace would soon be launched after it was postponed.

Williams also disclosed that Arweave 2.7 was being planned. Last year, the Arweave Protocol launched version 2.6 which brought a new block validation method. The method made Arweave mining more energy efficient, incentivized miners to store more data and generate a more efficient storage market.

“...I would say at this point, you know, like fixes that make the protocol more elegant. They don’t add more core features. They just improve slightly,” Williams said about Arweave 2.7

Adeola

Adeola is a journalist at Arweave News. As a former freelance journalist, his works were published by Newlines Magazine, The Continent and the Mail and Guardian. He has interest in the intersection of technology and human lives.

Sign up for newsletter

Sign up here to get the latest news and updates delivered directly to your inbox.