Here is the cleaned-up JSON response from the API this particular response was from a query 'Brave Search', and the 'extra_snippets' are extracted from this Wikipedia page. So, after doing a few queries with Brave's Search API - I was rather surprised to see how generous their snippets are in this example below - the 'extra_snippets' range from 150 to 260 words.

The effect of the use upon the potential market for or value of the copyrighted work.

The amount and substantiality of the portion used in relation to the copyrighted work as a whole.

The purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes.

It provides for the legal, non-licensed citation or incorporation of copyrighted material in another author's work under a four-factor balancing test:

Google's featured snippets tend to be rather short (no more than 50 words), which from a copyright point of view can be classified as fair use.įair use is a doctrine in the law of the United States that allows limited use of copyrighted material without requiring permission from the rights holders. An example of a typical Google Featured Snippet For its Data for AI product, Brave offers something called 'Extra alternate snippets', which are very similar to what we know as Google's Featured Snippets. Rather than talking about it too much, I thought the logical thing to do would be to sign up for the API and see what kind of data we can find. This one, Data for AI, lets you 'Feed results to AI models for inference', while their premium version of this same API lets you 'Cache/store data to train AI models' not only with 'regular' rights but also 'storage rights'. Brave offers numerous API products, some of which are specifically designed for AI. I think this story could have been fairly decent even without the copyright part, so before we get to the nitty gritty stuff - I can 100% confirm that Brave lets you ingest copyrighted material through their Brave Search API. Brave gives you 'rights' to use data for AI inference/trainingĪs you may have noticed, I used the word copyrighted for the title of this story.

An update from Brave has been added at the bottom of the article! ↗ I have published a new article which is an update and shows how Brave can scrape licensed content, read it here: Personally, I'm not a big fan of these conglomerates ingesting other people's work and then reselling it, which also leads me to the story I'm going to talk about today. Whether or not that is going to hold up in the long run, remains to be seen.Īfter Google published an announcement saying they're interested in exploring alternatives to robots.txt to provide broader control over AI-related content issues, I was curious to see what other search engines are doing in regard to AI, both for dealing with AI-generated content but also handling data. I'm fairly certain that I was not the only person in the world who thought to himself, 'Did they just yoink the entire Internet and bundle it together into a glorified copy and paste machine?' upon the release of ChatGPT.Īnd even though there are some concerns about the type of data that was used to train OpenAI's latest model, it seems that the overall stance of OpenAI and other companies working on similar projects is that it is fair use.

Search brave