Meta has removed a controversial artificial intelligence feature from Instagram that allowed users to generate modified images referencing public accounts without alerting the account holders. The feature, part of the company's new Muse Image generator developed by Meta Superintelligence Labs, was rolled out earlier this week alongside other AI tools. Following immediate backlash from users, talent agencies, and privacy advocates, Meta announced Friday that the feature is no longer available.
The feature in question worked by allowing individuals to @-mention public Instagram accounts in prompts for the Muse Image generator. The system would then reference the mentioned account's public photos to create new, AI-generated images. Crucially, the feature did not notify the referenced user that their images were being used in this way. This lack of transparency sparked concerns over consent, privacy, and potential misuse, especially given the history of AI-generated non-consensual intimate images targeting female celebrities and public figures.
The Rise and Fall of Muse Image's Referencing Feature
Meta introduced Muse Image as a flagship product of its dedicated AI unit, Meta Superintelligence Labs. The company promoted the referencing feature as a creative tool for generating personalized content. However, the rollout was met with swift criticism. TechCrunch published a guide on how to disable the feature, reflecting widespread concern among users. Within days, Meta reversed course. In a blog post, the company stated: “Our intent was to provide a useful creative tool and to give people control over whether their public content could be referenced in this way. We’ve heard the feedback that this feature missed the mark, so it’s no longer available.”
The decision to remove the feature was first reported by Puck News founding partner Dylan Byers, who noted that the move came amid scrutiny from users and talent agencies, including Creative Artists Agency (CAA). CAA and other agencies had reportedly raised alarms about the potential for the feature to be used to create unauthorized and exploitative images of their clients, many of whom have large public Instagram followings.
Context of AI Misuse on Social Platforms
Since the integration of generative AI with social media platforms, misuse has become rampant. One of the most troubling trends has been the generation of non-consensual nude images, often targeting female celebrities. Platforms like TikTok, Twitter (now X), and Meta’s own Facebook and Instagram have struggled to implement effective guardrails. While some platforms have introduced policies prohibiting such content, enforcement has been inconsistent, and tools designed to prevent abuse often fall short. The removal of Meta's referencing feature underscores the ongoing challenge of balancing creative AI applications with user safety and consent.
Meta has faced previous controversies over AI and image usage. In 2023, the company faced backlash for using public photos to train its AI models without explicit opt-in consent. The company later introduced a privacy center for users to manage data usage, but critics argued that the opt-out process was cumbersome and not widely advertised. The recent feature appears to have compounded these concerns by enabling direct, real-time generation of images based on specific public accounts.
Broader Implications for AI Governance
The incident highlights the growing tension between rapid AI innovation and the need for ethical guardrails. Meta Superintelligence Labs, formed in late 2025, has been tasked with accelerating Meta's AI capabilities across its suite of apps. However, this speed of deployment has occasionally led to missteps. The referencing feature was likely intended to showcase the generative power of Muse Image, but the absence of notification mechanisms and consent protocols made it vulnerable to abuse. Privacy advocates argue that such features should require explicit permission from referenced users, not just reliance on public account settings.
Talent agencies like CAA have become increasingly vocal about protecting their clients' digital likenesses. They have pushed for stronger legal protections and platform policies to prevent unauthorized use of images, particularly in the context of generative AI. The removal of the feature may set a precedent for how platforms handle similar tools in the future. However, it also raises questions about Meta's internal review processes. How did a feature with such obvious risks pass internal testing? The company has not provided details on its approval workflow.
Meta's decision to remove the feature is a acknowledgment that even with internal AI ethics boards, mistakes will happen. The company has previously published principles for responsible AI development, but critics note that these principles are not always reflected in products. The incident serves as a case study in the importance of proactive risk assessment and stakeholder consultation before launching features that interact with public data.
As Meta continues to expand its AI offerings, the company will need to balance innovation with trust. The Muse Image generator itself remains available, but without the referencing capability. Users can still generate images using other prompts, but cannot directly leverage public Instagram accounts. This limitation may reduce the appeal of the tool for some creators, but it also protects against the most obvious forms of abuse. Whether Meta can regain user trust after this episode remains to be seen. The company has not announced whether it will reintroduce a revised version of the feature with stronger consent mechanisms.
The broader tech industry is watching closely. Other platforms, such as X and Snapchat, have also experimented with AI image generation from public data, though with varying levels of guardrails. The backlash against Meta's feature may encourage a more cautious approach across the industry. Regulatory bodies in the EU and US have already begun scrutinizing AI-driven image manipulation, and incidents like this could accelerate calls for legislation requiring consent for use of personal data in AI training and generation.
Ultimately, the removal of the feature is a reminder that AI development must proceed with humility and responsiveness to user concerns. Meta's quick reversal suggests that the company values user trust enough to pull a feature even after investment, but the initial rollout indicates a gap between intention and execution. As AI becomes more embedded in social media, the lessons from this episode will likely influence product decisions for years to come.
Source: TechCrunch News