Meta Introduces Transparency Labels for AI-Generated Content

Share Us

Meta Introduces Transparency Labels for AI-Generated Content
04 May 2024
4 min read

News Synopsis

Meta, the parent company of social media giants Facebook, Instagram, and Threads, is implementing new measures to enhance transparency regarding AI-generated content. These changes, aimed at providing users with more information about the origin of content, signify Meta's commitment to addressing evolving challenges in the digital landscape.

“Made with AI” Labels Coming Soon 

Starting in May 2024, users will see a new label – "Made with AI" – on various types of content across Meta platforms. This update acknowledges the limitations of the current policy, which Meta describes as "too narrow."

Wider Scope of Labeled Content

The labeling will encompass a broader range of AI-generated content, including videos, audio, and images.

User Disclosure and AI Detection

The labels will be applied in two ways:

  • User Disclosure: If a user uploads content created with AI tools, they can choose to disclose this information, triggering the "Made with AI" label.

  • Meta's AI Detection: Meta will also employ its own system to detect "industry standard AI image indicators" within uploaded content. However, the specific details of this detection system remain undisclosed.

Shifting Focus: Context Over Removal

Meta is also revising its approach to moderating AI-generated content. This change is based on recommendations from the company's independent Oversight Board.

Changes in Moderation Approach

Contextual Moderation Approach:

In line with recommendations from the Oversight Board, Meta will adopt a contextual moderation approach for AI-generated content starting in July. Rather than outright removal, AI-generated content that does not violate community guidelines will be provided with additional context, promoting transparency and informed content sharing.

Prioritizing Existing Community Guidelines

Content that violates existing community guidelines, such as those addressing bullying, voter interference, and harassment, will still be subject to removal regardless of its creation method (AI-generated or not).

This update highlights Meta's commitment to providing context for AI-generated content rather than resorting to outright removal.

Meta's latest updates in 2024:

Transparency in AI-Generated Content (May 2024):

Shifting Focus in Content Moderation (July 2024):

  • Emphasis on Content, not Creation Method: AI-generated content will no longer be automatically removed solely because it was created with AI tools.

  • Prioritizing Existing Guidelines: Moderation will focus on the content itself. Material violating existing community guidelines on bullying, voter interference, and harassment will still be removed regardless of its creation method.

Additional Updates:

  • New Ray-Ban Smart Glasses with Meta AI and Video Calling (April 2024): Meta has released new styles of Ray-Ban smart glasses featuring video calling with WhatsApp and Messenger, along with capabilities powered by Meta AI for visual information retrieval.

  • Open Mixed Reality Ecosystem (April 2024): Meta is opening up the operating system for its Meta Quest devices to third-party hardware makers. This aims to facilitate the development of a broader range of mixed reality applications.

  • Meta AI advancements (April 2024): Meta introduced "Meet Your New Assistant: Meta AI, Built with Llama 3," highlighting advancements in its large language model technology.

News Not Yet Confirmed:

  • Facebook News Deprecation (April 2024 - US and Australia): In early April, Meta started deprecating Facebook News, the dedicated news tab, in the US and Australia. This follows a similar move in other countries and reflects a shift in focus towards short-form video content.

Finding More Updates:

TWN Special