Tumblr 和 WordPress.com 的母公司 Automattic 正在与 Midjourney 和 OpenAI 进行谈判，以为后两者提供用以训练 AI 的用户帖子数据。Automattic 称它已在 Tumblr 和 WordPress.com 上增加了一个设置，允许用户阻止人工智能公司抓取他们的帖子和网站。该公司称在法律上不必遵守这一点，因为到目前为止并没有法律禁止爬行数据。但 Automattic 承诺将尊重用户的偏好。此外，如果用户选择稍后退出，这些内容仍将从训练数据中删除。Automattic 希望通过出售数据获得新的收入，尚不清楚交易涉及的金额。另据 404 Media 报道，Automattic 还收集了许多不应公开的 Tumblr 帖子，包括某些私人、已删除和露骨内容，不清楚这些数据是否已发送给 AI 公司。
Automattic, the parent company of Tumblr and WordPress.com, is currently in negotiations with Midjourney and OpenAI to provide user post data for AI training. Automattic has stated that it has implemented a setting on Tumblr and WordPress.com that allows users to block AI companies from scraping their posts and websites. The company claims that it is not legally obligated to comply with this, as there are currently no laws prohibiting data scraping. However, Automattic has pledged to respect user preferences. Additionally, if users choose to opt out later, their content will be removed from the training data. Automattic hopes to generate new revenue by selling the data, although the exact terms of the deal are unclear. Furthermore, according to 404 Media, Automattic has collected a significant amount of Tumblr posts that should not be made public, including private, deleted, and explicit content. It is unclear whether this data has been shared with the AI companies.
微信扫一扫