Sana is AI (Artificial Intelligence) collaborative anchor designed and developed by India Today Group. Sana was first introduced to the public on Mar 18, 2023 by Kalli Purie, Vice Chairperson, India Today Group. Sana was claimed to be the India’s first AI anchor by its developer and first appeared on Aaj Tak news channel.
What is Sana AI Anchor technically?
Sana is an AI avatar with having Indian news anchor appearance with highly realistic human-like news anchoring capabilities. To make it clear, it’s an AI bot which is only available in its soft form and has no physical body like Humanoid robots. She has been programmed to look like a real person and appear on the news channel with almost human-like interaction with the audience.
Who developed Sana AI Anchor?
Sana was designed and developed by India Today Group, in India.
The technology used to develop Sana AI Anchor?
AI Avatars like Sana are created using a combination of technologies such as natural language processing (NLP), machine learning (ML), and text-to-speech (TTS) synthesis that makes it possible to create lifelike and interactive virtual characters.
Here’s a general overview of the process:
- Data Collection: Large amounts of text and speech data are gathered to train the AI model. This data could include conversations, books, articles, scripts, etc. The quality and diversity of this data are crucial for the AI to understand and mimic human speech patterns effectively.
- Natural Language Processing (NLP): NLP techniques are used to process and understand the textual data. This involves tasks like language modeling, sentiment analysis, entity recognition, and semantic parsing. NLP helps the AI understand the meaning and context of the text it encounters.
- Machine Learning Models: AI avatars often rely on deep learning models, such as recurrent neural networks (RNNs) or transformer models like GPT (Generative Pre-trained Transformer), to generate human-like responses. These models learn from the data collected in the previous step to generate text that is contextually relevant and coherent.
- Text-to-Speech (TTS) Synthesis: Once the AI has generated a response in text format, a TTS system converts this text into spoken words. TTS systems use neural network architectures to produce synthetic human-like speech. These systems are trained on large datasets of human speech recordings and learn to generate speech that sounds natural and expressive.
- Real-time Rendering: Facilitating live interactions, real-time rendering techniques are utilized, ensuring that avatars promptly respond and adapt to user inputs, delivering an engaging and dynamic experience.
Overall, creating AI avatars involves a combination of advanced technologies and iterative refinement to achieve human-like conversational abilities. While current AI avatars can produce remarkably realistic responses, there’s still ongoing research to enhance their capabilities further.
Note: We don’t have any official confirmation from India Today Group on the technology behind Sana but we can assure you that she is developed on the above tech stacks as the world of AI is being built on the same.
Capabilities of Sana AI Anchor?
Sana is an AI model which can do TV anchoring in many Indian languages having human-like anchoring capabilities.
Inabilities of Sana AI Anchor?
Sana is in its early development stage and lacks several human elements like human-like expressions, and facial emotions and still looks like an AI avatar.
Conclusion:
Overall, the future of AI avatars holds promise for revolutionizing human-computer interaction and enabling more natural, intuitive, and personalized experiences across diverse domains, from entertainment and education to healthcare and business.
