
Meta 微调了其商业条款,删去了暗示它可以自由用新闻网站(如 The Verge)训练 AI 模型的表述——但法律专家表示,这些改动不足以消除他们的担忧。
「这完全没有限制『与你的产品或服务相关』这类措辞,而 Meta 完全可以据此抓取整个网站,」康奈尔法学院数字与信息法教授 James Grimmelmann 对 Nieman Lab 表示。
最初,Meta 针对企业账户的新条款似乎赋予了公司在新闻网站上训练 AI 的许可。该条款 10 月 30 日生效,其中写道 Meta「可以使用」与企业账户关联的网站上的「任何内容」来「开发、研究」和「训练」Meta 的产品。当时就有法律专家告诉 Nieman Lab,这可能让 Meta「主张其商业条款优先于其他屏蔽指令」,例如长期以来告知抓取器(包括谷歌等搜索引擎)可以爬取哪些内容的 robots.txt 文件。
在出版业的强烈反弹之后,Meta 删去了将用抓取内容「训练」产品的表述。但这些微调可能还不足以排除这种可能性。「新版条款仍允许 Meta 用那些抓取的内容来开发任何『商业产品』。这似乎不是一个有定义的说法,而且 Meta 大概率正在开发 AI 商业产品。我认为『开发』的范畴足够宽,宽到可以包含『训练』,」Loevy & Loevy 律所的媒体与知识产权律师 Matt Topic 对 Nieman Lab 表示。
Meta 一直坚称新条款并非为了免费抓取新闻内容。公司发言人 Daniel Roberts 今天在给 The Verge 的声明中重申,「内容所有者可以用 robots.txt 表达他们的偏好」,并称此次更新「不是为了抓取新闻网站,这些更新也没有赋予我们凌驾于这些偏好之上的新权利」。
(信息来源:The Verge,2026-10-23)
本文来源:https://www.theverge.com/tech/1017082/meta-terms-update-ai-training-concerns
若有版权问题,请来信告知:13951431913#139.com 或在右方提交评论。
Meta has tweaked its business terms to remove language suggesting that it could freely train its AI models on news sites, like The Verge â but legal experts say these changes arenât enough to address their concerns. âThis does nothing to restrict the âassociated with your products or servicesâ [language] that would let Meta scrape entire sites,â James Grimmelmann, a professor of digital and information law at Cornell Law School, tells Nieman Lab.
Initially, Metaâs new terms for business accounts appeared to give the company permission to train its AI on news sites. The terms, which go into effect October 30th, stated that Meta âmay use any contentâ from the websites associated with a business account to âdevelop, researchâ and âtrainâ Metaâs products. At the time, legal experts told Nieman Lab that this could allow Meta âto argue its commercial terms supersede other blocking instructions,â such as the robots.txt file that has historically told scrapers (including search engines like Google) what they are allowed to crawl.
Following fierce backlash from the publishing industry, Meta removed language saying it would âtrainâ its products with scraped content. But the tweaks might not be enough to rule out that possibility. âThe revised version of the new terms still allows Meta to use that scraped content to develop any âbusiness product.â That doesnât seem to be a defined term, and presumably Meta is developing AI business products. I think âdevelopâ is broad enough to include âtrain,ââ Matt Topic, a media and intellectual property attorney at Loevy & Loevy, tells Nieman Lab.
Meta has maintained that its new terms arenât designed to scrape news content for free. In a statement to The Verge today, Meta spokesperson Daniel Roberts reiterated that âcontent owners can use robots.txt to signal their preferences,â adding that the update âis not about scraping news sites, and these updates donât give us new rights that supersede those preferences.â