Meta AI takes another step in the competition: its new image model Muse is fully integrated into Instagram and WhatsApp, and advertisers can also use it
Meta has taken another key step forward in its AI commercialization strategy.
On Tuesday, the 7th of Eastern Time, Meta officially launched its brand-new AI image generation model, Muse Image. This is the first image generation model released by Meta after a year of heavy investment in restructuring its AI team and establishing Meta Superintelligence Labs (MSL). The model will first be integrated into the Meta AI chatbot, and gradually rolled out across Meta's apps such as Instagram and WhatsApp. In the future, it will also be made available to advertisers to generate marketing materials.
What is even more noteworthy is that for the first time, Meta has deeply integrated Instagram social connections with AI image generation. Users can not only generate images from text and edit existing images, but also create relevant visuals based on public Instagram posts from friends or creators, making AI creation far more personalized. However, in response to copyright and privacy controversies, Meta has simultaneously introduced an "opt-out" option that prevents others from using their content for AI remixing, and added invisible digital watermarks to all AI-generated images.
Analysts believe that as companies including OpenAI, Anthropic, Google, and xAI continue to upgrade their image models, Meta's launch of Muse not only signifies a further expansion of its AI product portfolio, but also signals the company's ambition to fully embed AI capabilities into its social ecosystem, paving the way for future AI cloud services and enterprise-level commercialization.
01
Muse Makes Its Full Debut Across the Meta Ecosystem, Integrated into Instagram and WhatsApp
According to Meta, Muse Image will first be deployed in the Meta AI chatbot, and gradually extended to all of Meta's major products, including Instagram and WhatsApp.
Compared to Meta's earlier Emu model, Muse features a more comprehensive positioning, supporting: Text-to-Image generation; editing and modifying existing images; partial inpainting based on natural language; and future capabilities for generating marketing materials for advertisers.
Meta states that advertisers will soon be able to directly use Muse to produce ad creatives, product promotional images, and other marketing content. This means Muse will not only serve regular consumers, but also become a core component of Meta's AI commercialization efforts.
Industry insiders note that this is a significant move by Meta to further strengthen its AI advertising business. Advertising still accounts for the vast majority of Meta's revenue, and AI-generated marketing materials are expected to help small and medium-sized advertisers reduce production costs and improve ad delivery efficiency.
02
For the First Time, Instagram Social Connections Are Integrated, Allowing AI to "Generate Images of Your Friends"
The biggest product innovation of this launch is not image generation itself, but the deep integration with Instagram content.
Meta says users can ask the AI to generate a photo of a friend surfing at the beach, or place a creator in a futuristic city. The AI will reference the public photos these individuals have shared on Instagram to create the outputs.
In other words, public Instagram content will become a key reference resource for Muse, making generation results more consistent with real people, rather than relying solely on text prompts.
Of course, Meta is also aware that this capability could raise copyright and privacy concerns.
Therefore, Meta has simultaneously rolled out an opt-out mechanism. Any Instagram user can select in their settings to prevent others from using their public content for AI remixing or reuse.
This design is regarded as an important measure by Meta to address the generative AI copyright disputes that have emerged in recent years.
03
All AI Images Are Embedded with Hidden Watermarks, and Enhanced Safety Restrictions Are Implemented
In recent years, AI image generation has continuously sparked public controversy.
Since the start of this year, image models on other AI platforms have been misused to generate non-consensual explicit images and illegal content involving minors.
Meta states that Muse incorporates multiple layers of safety protection measures:
- All AI-generated images will be embedded with invisible digital watermarks
- The system will block generation requests that violate Meta's community policies
- Illegal content including child sexual abuse material (CSAM) is strictly restricted
Meta says the company is continuously improving its safety filtering mechanisms, aiming to enhance generation quality while reducing the risk of the model being misused.
04
The First Image Model One Year After the AI Lab Restructuring
Muse is also the first major achievement following Meta's restructuring of its AI R&D system.
Over the past year, Meta has carried out the largest organizational adjustment in its history centered on AI.
The company hired former Scale AI CEO Alexandr Wang at a high salary to lead its AI business, and established the new Meta Superintelligence Labs (MSL). At the same time, it recruited a large number of top researchers from companies including OpenAI and Anthropic, hoping to regain its competitive edge in AI.
In April this year, MSL launched its first large language model, and Muse is the first officially released image generation model from this team.
A Meta spokesperson also revealed that the company is developing a video generation model, which is expected to launch in the coming months to further enhance its AI multimodal capabilities.
05
The Next Step Targets AI Cloud Services, With Plans to Sell Model Capabilities to Developers
Behind the launch of Muse, Meta's larger strategic goal may not be image generation, but the commercialization of AI infrastructure.
Meta has disclosed that in the future, the company plans to provide AI model APIs including Muse to external developers through its cloud platform, allowing developers to access Meta's AI capabilities without deploying the models themselves.
This means Meta is gradually building an AI cloud service ecosystem similar to the OpenAI API, Anthropic Claude API, and Google Vertex AI.
Last week, reports stated that Meta is also exploring more cloud computing businesses, including: leasing AI computing power to enterprises, selling GPU computing resources, and providing full AI model hosting services.
Analysts believe this is a key path for Meta to improve the return on its massive AI capital expenditures.
In recent years, Meta has invested tens of billions of dollars in building AI data centers and purchasing a large number of GPUs. Over the past few months, Meta has also signed large-scale computing power cooperation agreements with companies including CoreWeave, Google, and Oracle to meet future model training and inference demands.
Meta says even with these measures, the company is still facing a shortage of computing power supply, so it will continue to push forward the construction of a new round of ultra-large-scale AI data centers.
06
From Catching Up to Commercialization, Meta Is Fully Completing Its AI Product Portfolio
Text-to-image generation has become a standard capability for almost all leading AI companies today.
OpenAI's GPT-Image, Google's Imagen, Anthropic's related capabilities, and Midjourney have all matured into fully-fledged products.
Although Meta previously had internal models such as Emu, and had licensed third-party technologies like Midjourney, it always lacked a unified image generation platform that could cover consumer use cases, advertising businesses, and future cloud services.
The launch of Muse means Meta has begun to build its own complete AI image product system.
More importantly, Muse not only serves the chatbot, but is also directly integrated into social platforms with billions of users such as Instagram and WhatsApp, and further extends to ad creation, enterprise APIs, and AI cloud computing businesses.
As Meta continues to expand its deployment of AI models, self-developed infrastructure, and commercial services, the company is gradually transforming from a social media platform into a comprehensive AI platform with consumer-grade AI products, enterprise AI services, and AI infrastructure capabilities.
This article is from the WeChat public account "Wall Street CN Max", written by Li Dan, and published by 36Kr with authorization.