Meta Expands Labels for AI-Generated Content to Include Photos and Audio Across Facebook, Instagram and Threads
Meta Expands Labels for AI-Generated Content to Include Photos and Audio Across Facebook, Instagram and Threads
Meta's existing manipulated media policy, written in 2020, only covered videos altered by AI to make a person appear to say something they didn't say.$Meta Platforms (META.US)$ has announced a significant policy shift regarding manipulated media on its platforms. The company will now be introducing labels for AI-generated content on Facebook, Instagram, and Threads.
$Meta Platforms (META.US)$ 宣佈了有關其平台上受操縱媒體的重大政策轉變。該公司現在將在Facebook、Instagram和Threads上推出人工智能生成的內容的標籤。
What Happened: Meta's new policy, announced on Friday, is a response to feedback from the company's Oversight Board. The board recommended that Meta broaden its approach to manipulated media and provide context through labels.
發生了什麼：週五宣佈的 Meta 新政策是對該公司監督委員會反饋的回應。董事會建議 Meta 擴大其操縱媒體的方法，並通過標籤提供背景信息。
The Mark Zuckerberg-led company has added photos and audio to its scope of coverage for this policy – only videos were considered before. However, it said it won't automatically remove AI-generated or manipulated content if it is not labeled – it will instead add the labels itself.
這家由馬克·扎克伯格領導的公司已將照片和音頻添加到該政策的覆蓋範圍中——以前只考慮過視頻。但是，它表示，如果沒有標籤，它不會自動刪除人工智能生成或操縱的內容，而是會自己添加標籤。
Meta's existing manipulated media policy, written in 2020, only covered videos altered by AI to make a person appear to say something they didn't say.
Meta 於 2020 年制定的現有媒體操縱政策僅涵蓋了人工智能修改的視頻，以使一個人似乎在說他們沒說的話。
The Oversight Board also advised against removing manipulated media that does not violate Meta's Community Standards, suggesting a "less restrictive" approach through labeling with context.
監督委員會還建議不要刪除不違反 Meta 社區標準的受操縱媒體，建議通過貼上背景標籤來採取 “減少限制” 的方法。
Meta has been working with industry partners on common technical standards for identifying AI content, and will introduce "Made with AI" labels on AI-generated video, audio, and images.
Meta 一直在與行業合作伙伴合作制定識別人工智能內容的通用技術標準，並將在 AI 生成的視頻、音頻和圖像上引入 “人工智能製造” 標籤。
The company plans to start labeling AI-generated content in May 2024 and will stop removing content solely based on its manipulated video policy in July.
該公司計劃在2024年5月開始爲人工智能生成的內容貼標籤，並將在7月停止僅根據其操縱的視頻政策刪除內容。
Meta says it will continue to keep this AI-generated content on its platforms and add informational labels and context unless it violates its policies.
Meta 表示，除非違反其政策，否則它將繼續在其平台上保留這些人工智能生成的內容，並添加信息標籤和背景信息。
Why It Matters: The decision to label AI-generated content comes at a time when the potential for AI to spread misinformation is a growing concern. The move is also in line with Meta's earlier efforts to introduce labels for AI-generated images across its platforms in response to the 2024 election year.
爲何重要：爲人工智能生成的內容貼上標籤的決定是在人工智能傳播錯誤信息的可能性越來越令人擔憂之際做出的。此舉也符合 Meta 早些時候爲回應 2024 年大選年而在其平台上引入人工智能生成的圖像的標籤。
The announcement also follows a series of events that have highlighted the potential for AI to exacerbate election misinformation, such as an AI-generated robocall featuring President Joe Biden that urged New Hampshire voters not to participate in the state's primary election.
該公告還是在一系列事件之後發佈的，這些事件凸顯了人工智能有可能加劇選舉錯誤信息，例如由喬·拜登總統發起的由人工智能發起的自動通話，敦促新罕布什爾州的選民不要參加該州的初選。
