一位 Mastodon 用户声称，AI 模型（特别提到了 Anthropic 的 Claude）未经同意就使用了他大量的博文和艺术作品进行训练。他认为，AI 现在写得像他一样，而不是反过来，并引用了 22 年的内容。该用户还表示，他在 Anthropic 问题上个人有切身利益，因为该公司据称非法使用了他撰写过章节的书籍进行训练，并因此成功起诉了该公司。 AI
影响 凸显了对 AI 训练数据来源和版权侵犯的持续担忧，可能影响未来的政策和法律挑战。
排序理由 社交媒体平台上的用户生成内容，表达了关于 AI 训练数据和过往法律纠纷的意见和不满。
在 Mastodon — mastodon.social 阅读 →
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
So if anyone goes 'hey you write like AI' - well guess what... I know they trained off my artwork, it's more than likely they trained off my 22 years of blogging as well at Radio Clash and my porfolio sites wherre they found the art, and various blogs. Many 100,000s of words ther…
I have skin in the game re: Anthropic - they trained illegally off a book I wrote a chapter in - Podcasting Hacks. And got successfully sued. I now find they are chopping books to bits, scanning them and trashing them, even rare ones, well fuck them. https:// arstechnica.com/ai/2…