Meta unveils new strategies to combat AI-generated content

Meta, the global social media giant, has recently announced pioneering changes to its policies concerning AI-generated content and manipulated media. These alterations come as a response to critiques from Meta’s Oversight Board, aiming to confront issues related to deceptive information and misinformation.

Commencing next month, Meta will introduce a distinguishing ”Created with AI” emblem for deepfakes, which are altered media produced using artificial intelligence. Furthermore, in cases where media has been manipulated in ways that could mislead the public on critical matters, Meta will provide users with additional contextual information.

This policy adjustment carries significant implications, particularly in a year marked by numerous elections worldwide. It signifies that Meta may classify more content with the potential to misinform, ensuring transparency for its users. However, the labeling of deepfakes will be constrained to content exhibiting ”industry standard AI image indicators” or identified by the uploader as AI-generated.

While this shift may result in more AI-generated content and manipulated media persisting on Meta’s platforms, the company believes that transparency and supplemental context offer a better approach to managing such content instead of outright removal.

In line with this modification, Meta will cease removing content solely based on its existing manipulated video policy in July. This timeline aims to familiarize users with the self-disclosure process before content ceases to be removed.

This updated approach likely stems from the mounting pressure encountering Meta concerning content moderation and systemic risks. The implementation of the European Union’s Digital Services Act, which imposes regulations on Meta’s social networks, necessitates the company to strike a delicate balance between eliminating illicit content and protecting freedom of speech.

The critical feedback from the Oversight Board played a significant role in Meta’s decision to reevaluate its policies. With the Board’s scrutiny emphasizing the inconsistency in Meta’s policy, which previously only addressed AI-created videos, leaving other forms of manipulated media unaddressed, Meta intends to expand its scope to encompass AI-generated audio, images, and other forms of manipulated media. By utilizing signals endorsed by the industry to identify AI content, users will have the opportunity to disclose whether their uploads contain AI-generated material.

In scenarios where digitally fabricated or altered media poses a substantial risk of deceiving the public on vital issues, Meta may employ more visible labels to furnish additional information and context. The objective is to equip users with the essential tools to evaluate content accurately and comprehend its context across diverse platforms.

It is of paramount importance to acknowledge that Meta will not eradicate manipulated content, whether AI-based or otherwise, unless it contravenes other policies such as voter interference or harassment. Instead, the company will append informational labels and context within situations generating heightened public interest.

For content accuracy assurance, Meta has collaborated with close to a hundred independent fact-checkers to aid in the identification of risks associated with manipulated media. These external entities will remain active in reviewing counterfeit and misleading AI-generated content. Upon categorizing content as ”False or Altered,” Meta will implement algorithmic modifications to diminish its reach and exhibit an overlay label equipped with more information for users encountering it.

As Meta strides to tackle concerns concerning AI-generated content and manipulated media, it strives to maintain a balance between combatting misinformation and safeguarding the principles of free speech.

The source of the article is from the blog jomfruland.net