Platform Governance Archive

Author: Ayşe Darakçı

  • Guest Lecture by Prof. Tim R. Samples on The Rules of Digital Platforms: Introducing the 1999-2025 LDT Corpus

    Terms of use (TOUs) are among the most prolific contracts in human history, governing digital rights at societal scale. Yet their evolution over time is less examined, and public data is scarce. As part of the Media Structures and Platform Governance seminar, on 26 May 2026, Prof. Tim R Samples (University of Georgia) joined us to introduce…

  • DeepSeek Holds Developers Responsible for End-User Compliance

    Effective March 27, 2026, new developer terms require developers to ensure end-user compliance with DeepSeek’s terms, extending liability beyond the developer’s own usage (line 35). Regarding disputes between the developer and end user, DeepSeek assumes no responsibility. Developers bear full liability for any end-user breaches of the terms. (line 77-85/R) Additionally, DeepSeek can request proof…

  • WeChat Drops the Word ‘Permanent’ from Deletion Promise

    Effective 07.04.2026, WeChat updated its Privacy Policy. The new update drops the word “permanent” in various data deletion practices such as Chat Messages & Chat Media (line 161), Log Data (line 163) and Location Data (line 164).The removal of the word “permanently” may imply that WeChat do not guarantee irretrievable erasure of user data from…

  • Spotify Can Allow Prediction Market Advertisements

    On 4 March 2026, Spotify did a small but interesting update. The new Spotify, Advertising Content Policy update lists “prediction markets” aka betting markets or gambling on real world events under the examples of ads that they may permit in certain instances. (line 98/R). Author: Ayşe Darakçı 🔗 Link to the version: https://github.com/OpenTermsArchive/pga-versions/commit/f17d979edd4b6bd50d03cdc7c8b52bc77e32f463?diff=unified

  • Meta is Taking Measures to Prevent Scam Ads on Facebook and Instagram

    On March 23, 2026, Meta updated its Advertising Content Policy, indicating that advertising accounts must be associated with a legitimate business or individual. Last October, a report from Investigate Europe showed how scammers exploit EU regulatory loopholes and the lack of platform liability that enables, for instance, the spread of financial advertisements using celebrity deepfakes.…

  • Mistral AI Ease Up Data Usage: User Data and Search Results Now Used for Training

    Effective March 12, 2026, the new update from Mistral AI’s Commercial Terms states that customer data and outputs can now be used to train its AI models for free subscribers, Le Chat Pro users (unless opted-out), and Labs Model users. (line 79) Labs Models are introduced as experimental pre-release models exclusive to commercial customers. Data from…