0
0
Meta’s Updated Policy on AI-Generated Content Labeling
Meta recently revealed new policy adjustments aimed at ceasing the suppression of innocuous AI-generated content, shifting towards a strategy of categorizing a broader spectrum of video, audio, and image material as ‘Made with AI.’
Decisive Motivations Behind the Policy Updates
These modifications emerge as a direct response to Meta’s conscious choice not to eliminate a contentious post that depicted President Joe Biden in a compromising scenario with his granddaughter, accompanied by an offensive label branding Biden as a “pedophile.” The Oversight Board supported Meta’s stance on retaining the post online while critiquing the existing manipulated media policy for being excessively restrictive, illogical, and perplexing for users.
Previous Limitations Addressed by the Oversight Board
Formerly, Meta’s guidelines dictated the removal of “videos edited or produced by AI to fabricate statements from an individual.” The Oversight Board highlighted the shortcomings of this approach, failing to encompass other forms of manipulated media such as “cheap fakes,” altered audio, or content portraying individuals engaging in actions they did not commit.
Enhancements to the Labeling Process
Monika Bickert, Meta’s Vice President of Content Policy, acknowledged the Oversight Board’s critique in a blog post, acknowledging the necessity to broaden the scope of AI-manipulated content addressed through the implementation of the “Made with AI” classification. This new policy will involve affixing such labels to identified AI-generated content as well as content voluntarily disclosed as AI-created.
Embracing a Less Intrusive Approach
Meta’s Oversight Board emphasized the negative repercussions of removing AI-generated content that did not directly breach community standards, potentially encroaching upon freedom of expression. Consequently, Meta has decided to adopt a less prohibitive strategy by introducing labeling measures rather than content censorship.
Future Roadmap Towards Greater Transparency
Commencing in May 2024, Meta’s labeling initiatives will aim at furnishing users with added context and transparency regarding AI-manipulated content. Moreover, come July, Meta will halt the removal of AI-generated content falling outside the confines of specific regulatory guidelines.
Crucial Policy Consolidation Following Oversight Board Recommendations
Lastly, Meta has taken steps to amalgamate its manipulated media policies in line with the Oversight Board’s recommendations, expressly outlining the harms it aims to prevent beyond mere misinformation. This inclusive approach seeks to fortify public awareness and context surrounding varying forms of manipulated media.
Image/Photo credit: source url
