Alibaba Group Holding Ltd. is stepping into the competitive AI landscape with its latest innovation, an artificial intelligence model designed to interpret human emotions. This new model, R1-Omni, developed by the team at Alibaba’s Tongyi Lab, not only deciphers emotional states from videos but also provides detailed descriptions of individuals’ attire and surroundings, elevating the realms of computer vision.
In a move to solidify its position in the AI sector, the e-commerce giant looks to outshine competitors like OpenAI. R1-Omni, an enhancement over the HumanOmni model by lead researcher Jiaxing Zhao, is now available for free download on Hugging Face, underscoring Alibaba’s strategy of leveraging open-source platforms to foster innovation.
This initiative is a part of Alibaba’s broader strategy following the attention-grabbing entrance of DeepSeek in the AI space earlier this year. The company is rolling out various AI innovations across diverse sectors, having already tested its Qwen model against DeepSeek and struck a partnership with Apple Inc. for AI integration on iPhones.
The concept of emotionally intelligent AI, where systems can interpret and react to human sentiments, is increasingly being embraced. Such technologies are evident in customer service chatbots that identify user frustration and Tesla Inc.’s vehicles that detect driver fatigue.
In China’s aggressive market, Alibaba is waiving fees for its latest AI advancement, enabling unrestricted access to R1-Omni. While current demonstrations depict the model highlighting broad emotional states like “happy” or “angry,” its ability to extract these from visual contexts signifies a notable leap in AI capabilities.
Earlier this year, Alibaba had announced that its video generation artificial intelligence models can be used without cost. These models are the four open source models that are part of Wan2.1 series. Academics, researchers and commercial institutions around the world can now access them through Alibaba Cloud’s Model Scope and Hugging Face to generate images and video from text and image inputs for free.
Ever since China’s DeepSeek AI, which is also an open source model like Alibaba’s, was released in January with the claim of using lower cost and less advanced chips, people around the world have been paying attention to open-source AI tech.
Open source AI is an AI system including code, algorithms, and datasets, that the public can download and modify freely. Its purpose is not to generate revenue but to accomplish other purposes like creating the community or improving the product.
Alibaba released its first open source model in August 2023, and now is one of the most popular AI in the world, alongside DeepSeek’s. Its share in Hong Kong this year went up 66% due to factors like having better financial performance, being a key AI player in the nation, and possibly receiving support from its president, Xi Jinping, for the domestic private sector.
