Meta removes AI feature from Instagram after privacy backlash

Meta has removed its artificial intelligence feature 'Muse Image' from Instagram and the Meta AI app just three days after launch, following a wave of criticism over privacy violations. The tool, which allowed any user to modify photos from public Instagram accounts using generative AI, did not include any notification to the owner of the original image. The company acknowledged that the design 'did not meet privacy expectations' and that the feature 'missed the mark.'

According to Puck News, the decision came after scrutiny from users and talent agencies, including the Creative Artists Agency (CAA). Meta issued a statement on Friday announcing the removal of the feature, which had been launched the previous Tuesday under the direction of Alexandr Wang, head of AI at Meta Superintelligence Labs.

What was Muse Image and why did it cause controversy?

Muse Image was the first image generator from Meta Superintelligence Labs. It allowed users to take any public Instagram photo and modify it using text commands, generating new versions of the image without the consent of the original author. The lack of automatic notification was the trigger for criticism, as users could see their photos altered without knowing it, raising concerns about unauthorized use of personal images and potential malicious uses such as deepfakes or harassment.

The controversy quickly intensified on social media, with content creators and celebrities expressing outrage. The CAA agency, which represents numerous public figures, pressured Meta to remove the feature, arguing it violated image rights and exposed their clients to reputational risks.

Implications for privacy and AI regulation

This incident underscores the regulatory challenges tech companies face when implementing generative AI on platforms with user-generated content. Although Meta argued that public photos were already accessible, the ability to modify them without consent represents a new level of intrusion. Privacy experts note that the feature could have violated the General Data Protection Regulation (GDPR) in Europe and similar laws in other jurisdictions, which require explicit consent for processing personal data, including images.

The swift removal of Muse Image contrasts with the approach of other companies like OpenAI and Google, which have implemented stricter safeguards in their image generators. For example, OpenAI's DALL-E 3 includes content filters and does not allow generating images of real people without consent. Google Imagen also has similar restrictions. Meta, on the other hand, launched the feature without these protections, drawing criticism over its internal review process.

Consequences for Meta and the future of AI on Instagram

The removal of Muse Image is a setback for Meta's plans to integrate generative AI into its platforms. The company had heavily invested in Meta Superintelligence Labs and in hiring Alexandr Wang, former head of AI at Scale AI. This failure could delay the launch of other similar AI features and increase regulatory scrutiny on Meta.

For Instagram users, the lesson is clear: public photos can be used by AI without their knowledge, even if the feature is temporarily removed. It is recommended to review privacy settings and consider making accounts private to avoid future unauthorized use. Additionally, this case could drive greater regulation of generative AI on social media, similar to deepfake laws already in place in some countries.

What should readers know?

This incident demonstrates that generative AI is advancing faster than social and legal norms. Users should be aware that their public data (photos, texts, etc.) can be used to train or feed AI models without their explicit consent. Platforms like Instagram should implement clearer notifications and controls. Meanwhile, pressure from users and agencies can be effective in curbing invasive features, as seen in this case.

“Meta has acknowledged that the feature 'did not meet privacy expectations,' but the question remains: why was it launched without those safeguards?” — TheVortiq

In conclusion, the removal of Muse Image is a temporary victory for privacy, but it underscores the need for stronger regulation and better practices in AI development. Meta will need to reconsider its strategy if it wants to avoid future missteps.