Meta unveils Muse Image, its inaugural AI image model developed within Alexandr Wang's lab.
TL;DR: Meta has introduced Muse Image, an image-generation AI model across Meta AI, Instagram, and WhatsApp. Users can create images of friends from public posts, with advertisers gaining access soon. An opt-out option is available.
On Tuesday, Meta unveiled Muse Image, its inaugural image-generation AI model developed under the Superintelligence Labs, headed by Chief AI Officer Alexandr Wang. This model is being integrated into the Meta AI chatbot and will also be featured on Instagram and WhatsApp. Users can create images from text prompts or modify existing photos.
A key feature allows users to generate images that include friends or creators based on their public Instagram posts. Users who prefer not to have their content utilized or remixed by the AI can opt out through the settings menu. All images generated with Muse Image will bear an invisible watermark, and the system is equipped with safety measures to ensure compliance with Meta’s terms of service, including preventing CSAM.
Advertisers will soon be able to use the model for marketing material creation. In April, Meta's Superintelligence Labs had already introduced its first large language model, Muse Spark, and is also developing a video generation model anticipated to launch in the coming months. This image model represents the second significant product from the lab, with Wang having been brought on board a year ago following Meta’s multi-billion dollar investment and the hiring of several researchers with compensation packages over $200 million.
Meta intends to offer access to its AI models to external developers through a cloud service, enabling the company to monetize its data center resources and costly AI chips. This cloud initiative is part of a larger strategy reported by Bloomberg last week, which may involve selling raw computing power. Nevertheless, Meta still seeks additional computational resources, recently securing partnerships with CoreWeave, Google, and Oracle. The company made 8,000 job cuts in May and is redirecting billions toward AI infrastructure as it strives to compete with OpenAI and Anthropic in all areas of AI.
Other articles
Apple may have discovered an innovative method to enhance the battery life of the iPhone Air 2.
The upcoming slim iPhone from Apple might finally achieve a balance between style, battery life, and improved cameras while remaining thin.
Microsoft replaces OpenAI's technology with its own AI in certain applications.
Microsoft is shifting certain app features to its own MAI models rather than relying on OpenAI and Anthropic, reducing its dependence on external partners.
OpenAI is seeking to hire an investment banker for its AI division.
OpenAI's $205K position for a "subject matter expert" intends to instruct its AI in the routine tasks of Wall Street while competing with Anthropic for profitable revenue in the finance sector.
Apple might be increasing the thickness of the iPhone 18 Pro to accommodate a significant camera enhancement.
The upcoming Pro iPhone from Apple may feature a notably thicker design, probably because of a larger battery and a speculated variable aperture camera.
Microsoft replaces OpenAI with its own AI in several applications.
Microsoft is directing certain app features to its own MAI models rather than relying on OpenAI and Anthropic, reducing its dependence on external partners.
The prohibition of social media for young users is turning into a significant challenge for age verification.
Researchers set up 50 social media accounts that purported to be from 16-year-olds and discovered that none of the leading platforms required users to verify their age.
Meta unveils Muse Image, its inaugural AI image model developed within Alexandr Wang's lab.
Meta's Muse Image feature is now available in the Meta AI chatbot, Instagram, and WhatsApp. Users have the ability to create images of friends based on public Instagram posts. There is an option to opt out.