Show HN: Changewatch – see what’s changed between versions of policies

This project came about because we kept getting emails from companies saying something along the lines of “We’ve updated our terms of use. Click here to see the new version”. This was very annoying, as most of these emails would, at best, include a summary of the changes, but most often not give any indication of what changed, leaving you to hope there’s an archived version you can compare the new version with. We made Changewatch to help people see exactly what’s changed between versions of policies in a simple, human-readable way, so that they can know precisely what changes they’re agreeing to.

We’ve started by monitoring things like privacy policies from large tech companies like Google, Apple, Amazon, etc., and over time we’ll expand the number of policies we track. Changewatch currently only accesses the UK versions of policies, but we’re planning to support multiple locales per policy in a future update.

Sites like https://tosdr.org/ do a good job of breaking down the meaning of different policies, but don’t solve our problem of knowing that something we had agreed to had changed, but not precisely what had changed. Some companies, like Google, keep archived versions of old policies, and include diffs, but this is by no means the norm, and in those cases Changewatch provides an independent verification of their claims.

We experimented with several ways of diffing natural-language text to try and find one that yields an intuitive and readable diff display. We ended up with a two-stage process where we first compare the two pieces of text by paragraph, and then by word. The resulting diff matches nicely with the original text structure, which helps with readability and comparison with the original document.

We faced several challenges with the scraping, some of which aren’t solved yet. For instance, Facebook (seemingly) randomly serves our scraper either the UK or the EU version of their privacy policy, even though we’re accessing it via a UK-based proxy. This results in policy revisions being created where the only changes are the differences between the UK and EU versions of the policy. Hence why Changewatch doesn’t currently track Facebook’s privacy policy.

Thanks for reading. We would love to hear your comments, feedback and suggestions.