[DigitalToday reporter Jinju Hong (홍진주)] Meta is applying its new artificial intelligence (AI) model, Muse Spark, across its platforms as it moves to shift its AI strategy.

On April 9, IT outlet TechRadar reported the results of using Muse Spark. It said it felt closer to a "ChatGPT-like experience" inside social platforms than a simple auxiliary feature.

Meta applied Muse Spark to the Meta AI app and web services in the United States. It is pitching it as the default AI experience rather than a separate feature. It plans to expand the rollout to Instagram, Facebook, Messenger, WhatsApp, Meta AI glasses and other services.

The biggest change is its role. Where Meta AI previously stayed at the level of sticker creation or caption help, Muse Spark has evolved into something closer to a conversational assistant. The web interface has also shifted to a chat-window-centered layout like ChatGPT. It supports file uploads and image and video generation. In actual use, the natural flow of conversation and ability to carry out tasks stand out, giving the impression that its usefulness has increased significantly.

On performance, multimodal processing and handling of constraints were confirmed. For example, it produced results that maintained conditions even for complex requests such as writing lyrics that exclude specific letters of the alphabet and generating audio. The completeness was somewhat disappointing in parts, but its ability to build content while keeping constraints was sufficiently demonstrated.

Another strength Meta highlighted is combining internal platform search with character-style responses. It can generate answers using so-called "social context," such as reflecting reactions to specific topics on Facebook or constructing conversations with character settings. A key feature is that it naturally extends beyond providing simple information to generating scenario-based content.

Image generation quality is also relatively stable. The finishing touches were clean and the accuracy of text rendering stood out. It was assessed as being suitable for producing social content such as memes or posts. Coding functions were also added, supporting code writing, debugging and explanations in various languages. It does not yet support background tasks such as running a timer.

Meta's strategy is not only about competing in the standalone AI app race. The key is to embed AI as if it were the default across its platforms used by billions of people. Even if Meta AI offers an experience similar to ChatGPT or Gemini, being used naturally within social feeds and messaging flows works as a differentiating factor.

Against this backdrop, Muse Spark shows Meta's AI competition strategy shifting from adding features to making it the platform default. A forecast has emerged that even if Meta fails to take the No. 1 spot in the chatbot market, the competitive landscape could change if it anchors AI first in everyday scrolling, conversations and content creation.