Meta Takes This Step To Combat Deepfakes, New Measure To Be Effective From May

Meta, the parent company of Facebook and Instagram, has announced its decision to introduce labeling for AI-generated media starting in May. This measures aims to address concerns surrounding deepfakes and reassure both users and governments.

In response to criticism from its oversight board, which independently reviews content moderation decisions, Meta will implement changes to its approach to manipulated media. Rather than removing such content outright, Meta will now rely on labeling to preserve freedom of speech.

AI technology has made it easier to create convincing deepfakes

The decision follows urgent calls from the oversight board to revamp Meta's strategy in light of significant advancements in AI technology, which have made it easier to create convincing deepfakes. This move comes amidst fears of widespread misuse of AI-powered applications for spreading misinformation, especially during pivotal election periods globally.

Under the new policy, Meta will introduce "Made with AI" labels to identify content created or altered using AI across various media formats, including video, audio, and images. Content deemed to pose a high risk of misleading the public will receive a more prominent label.

There may be limitations, caution experts

While acknowledging the significance of identifying AI-generated content, some experts caution that there may be limitations, particularly with open-source software that may not adhere to the agreed watermarking standards.

The rollout of Meta's labeling initiative will occur in two phases, with AI-generated content labeling set to commence in May 2024. Additionally, the removal of manipulated media solely based on the old policy will cease in July, in accordance with the new standard.

Despite these measures, concerns persist over the proliferation of convincing AI deepfakes. Recent incidents, including manipulated videos of US President Joe Biden and AI-generated speeches by political figures like Imran Khan in Pakistan, underscore the urgency of addressing the risks associated with AI manipulation of media.