Elon Musk promotes algorithmic transparency as Twitter makes source code public
1 min read 01 Apr 2023, 07:33 AM ISTTwitter has made its recommendation algorithm's source code public by uploading it to two repositories on the code-sharing platform GitHub.
Twitter has made parts of its recommendation algorithm's code public, making users and programmers understand how the platform recommends content and suggests modifications to the algorithm.
Twitter has made its recommendation algorithm's source code public by uploading it to two repositories on the code-sharing platform GitHub.
This release includes the source code for many parts of Twitter, including the recommendations algorithm that determines the tweets users see on their timelines.
Musk believes that code transparency would increase user trust and lead to rapid improvements in the product. The move is also a response to concerns from users and lawmakers who are scrutinizing social media platforms over how algorithms select content for users.
Musk tweeted on Friday that third parties should be able to analyze the open-sourced code and "determine, with reasonable accuracy, what will probably be shown to users."
Elon Musk has reportedly stated that Twitter will update its recommendation algorithm based on user suggestions every 24 to 48 hours.
This move would allow the company to address concerns and make rapid improvements to the algorithm in response to user feedback, as reported by Reuters.
On Friday, Musk and some Twitter employees held a session on Spaces, Twitter's audio chat feature, where they invited users to bring recommendations and questions about how the platform's code works.
Twitter's repositories on Github do not include the code that powers the platform's ad recommendations.
The company has clarified that it has excluded this code from the public release. Additionally, Twitter has stated that it excluded any code that would compromise user safety or privacy, as well as details that would undermine efforts to prevent child sexual abuse material on the platform.