Meta has unveiled significant changes to its policies regarding AI-generated content and manipulated media, following critique from its Oversight Board.
Commencing next month, the company will implement broader labeling for such content, including the introduction of a “Made with AI” badge for deepfakes. Additionally, Meta will provide additional contextual information for content manipulated in ways that pose significant risks of deceiving the public on critical issues.
This shift in policy may result in Meta labeling more content that has the potential to mislead, which is particularly crucial in a year marked by numerous elections globally. However, for deepfakes, Meta will only apply labels to content with “industry standard AI image indicators” or when the uploader discloses that it’s AI-generated.
Consequently, AI-generated content falling outside these parameters may remain unlabeled. The company aims to adopt an approach focused on transparency and context, prioritizing this over removing manipulated media, due to associated risks to freedom of speech.
Meta has announced it will cease the removal of content based solely on its current manipulated video policy starting in July. This timeline allows users to understand the self-disclosure process before the removal of a smaller subset of manipulated media.
This policy change likely stems from increasing legal demands on Meta concerning content moderation and systemic risks, including regulations like the European Union’s Digital Services Act. Meta is under pressure to strike a balance between purging illegal content, mitigating systemic risks, and safeguarding free speech.
Policy Process Informed By Global Experts and Public Surveys
A majority of stakeholders agreed that removal should be limited to only the highest risk scenarios where content can be tied to harm, since generative AI is becoming a mainstream tool for creative expression.
This aligns with the principles behind our Community Standards – that people should be free to express themselves while also remaining safe on our services.
We also conducted public opinion research with more than 23,000 respondents in 13 countries and asked people how social media companies, such as Meta, should approach AI-generated content on their platforms. A large majority (82%) favor warning labels for AI-generated content that depicts people saying things they did not say.
The Oversight Board, which advises Meta independently, recently recommended amendments to its approach to AI-generated content. Meta has agreed to these suggestions, acknowledging the need to address a broader spectrum of manipulated media beyond AI-generated videos.
Meta’s engagement with independent fact-checkers aims to identify risks associated with manipulated content. These fact-checkers will play a crucial role in reviewing false and misleading AI-generated content, guiding Meta’s response, and reducing the reach of such content through algorithmic adjustments.
As synthetic content continues to proliferate, Meta’s policy shift underscores the growing need for enhanced transparency and context in combating misinformation and safeguarding public discourse on its platforms.