This is my first post on Paragraph. I'm just wondering how 'immutable' the posts are. If they are, some will say that it is great because they are censorship-resistant. But others will complain that it violates the 'right to be forgotten' and makes it hard or impossible to correct mistakes after publication. Does the web3 world take into account important legislations like GDPR, or do they think a certain pseudo-anonymity puts them above the law? Let's find out. Here is a step-by-step account of my first experiences.

First step: So I published the post, using the option to put it on Arweave. "Arweave is a decentralized data storage technology that allows information to be stored permanently and securely on a kind of 'collective hard drive' shared by many users around the world. Unlike traditional storage systems where data can be modified or deleted, information stored on Arweave remains immutable, meaning it cannot be changed once it is recorded," ChatGPT helpfully explains to me.

This post is very innocent, but suppose I wrote some fake news about someone. Even if I wanted to, or even if I were ordered by a court, it would be impossible to change or delete the post, even though GDPR legislation requires this to be possible.

Second step: Surprise! I checked and my edits have been applied. So is my post mutable or immutable? I decide to "collect" the post on Base, which is a second layer solution on the Ethereum blockchain. That is free, but there is a small fee of 0.000777 ETH. But it doesn't change my possibilities to edit.

Third step: While on the surface everything seems to be okay, in the sense that I still have the possibility to edit and delete the post on the blog, the situation is more complicated when I look at Arweave. I use the playground https://arweave.net/graphql to find the current and earlier versions of this blog post, and there they are! This means that anyone with the transaction IDs can find earlier versions. Or, if they don't have the IDs, they can search on Arweave for content based on keywords, tags, or addresses and find the earlier versions. Bye bye 'right to be forgotten'.

Conclusion: There are some ways to mitigate this, like using encryption, but that's not very handy if you want to reach a large audience. I think that for now the only way to handle this situation is to be very prudent when referring to other people in any post that ends up on Arweave. The good thing is Paragraph makes Arweave opt-in. One can only hope that users of this service think long and hard before putting data immutably on Arweave.