Key Dates
11 Dec -Extended Deadline
14 Dec -Judging
17 Dec -Winners Announced
Image Credit : Image Source Statement
All accompanying product images and usage scenario photographs were captured by the project team. The images represent the actual design, form, and real-world application of Kidodo, and are not conceptual renderings or stock imagery.
Project Overview
Kidodo is an AI-powered learning companion robot designed for children aged 0–10, aimed at becoming a child’s “smart growth partner.” With multimodal interaction as its core design philosophy, Kidodo can “see, hear, and respond” to children, helping them learn languages, develop reading habits, and expand their cognitive boundaries through natural conversation.
Its design features soft, rounded shapes made with food-grade silicone and an interchangeable "wearable system," creating a warm, friendly, and lifelike emotional experience. Kidodo uses an end-to-end real-time multimodal model, enabling high-precision visual recognition and natural speech interaction. It supports 23 languages, including English, Chinese, Japanese, and Spanish, and can recognize and explain picture books, magazines, doodles, and real-world objects.
With an innovative private-first architecture powered by PrivateLoRA technology, Kidodo ensures data security and privacy. It also features recyclable materials and modular construction, promoting sustainability and extending the product's lifecycle.
Kidodo blends cutting-edge AI with emotional design to create a truly unique and safe experience for children and parents alike.
Organisation
AUTOARK INTELLIGENT TECHNOLOGY (HONG KONG) CO., LIMITED
Team
Core Team: - Dr. Zeng Xiaodong, Founder Dr. Zeng, a disciple of internationally renowned computer scientist Isabel Tracoso, is one of China’s early pioneers in machine translation. He was selected for the MIT TR35, a list of the top 35 innovators under 35, which also includes notable AI entrepreneurs such as Kimi Yang Zhilin, Meituan’s founder, and Kuaishou’s Su Hua. With over 10 years at Alibaba and Ant Group, Dr. Zeng led the core engine of Alibaba's machine translation system. He is also a co-founder of Ant Group's Tech Lab, where he spearheaded innovative products like “unmanned supermarkets,” “facial recognition payments,” and “IoT hardware.” - Yang Dongyun, CPO A graduate of the University of Edinburgh, home to 20 Nobel laureates, Yang has won prestigious design awards such as the Red Dot and IF Design Awards. With extensive experience at Huawei, Alibaba, and ByteDance, she has held roles as product manager, design director, and business lead. Yang has led cross-functional teams, including product, user research, design, and growth. Her projects have won accolades such as the Harvard Business Review China New Growth Digital Innovation and Global Golden Trend Empowerment Innovation awards. She is also an active speaker at conferences like PCon Global Product Innovation and Bund Summit and holds over 20 patents. - Alex, Chief Architect A serial entrepreneur and multiple national champion in AI algorithm competitions, Alex holds a Master’s degree from Tongji University and the Hong Kong Polytechnic University. He specializes in building AI-native data intelligence systems using big data, NLP, and knowledge graphs. - Vincent, Model Scientist A Master's graduate from Fudan University, Vincent has published papers at top-tier conferences like ICML and KDD. His research expertise lies in machine learning and data modeling. - Jay, Product Director Jay holds dual degrees in Information Science and Architecture from Cornell University, leading projects that have won numerous international awards. He has successfully incubated new businesses from 0 to 1, expanding them into global markets. - Joe, Business Director With over 10 years of experience in marketing and business management, Joe has won multiple "Outstanding Management" and "Best Team" awards. He has managed PR and go-to-market (GTM) strategies for tech companies listed on the A-shares and US stock exchanges. Joe also has rich experience in smart healthcare and IoT industries.
Project Brief
Project Summary
Kidodo is powered by a multimodal real-time interaction model that integrates computer vision, speech understanding, and emotion recognition. It can perceive a child’s facial expressions, tone of voice, and emotional cues, responding in a natural and empathetic manner.
Kidodo transforms everyday life into a learning environment. It can read aloud any physical picture book, magazine, or real-world object, converting daily moments into immersive educational experiences. Supporting learning in 23 languages, it offers bedtime storytelling, emotional conversations, and personalized reading guidance tailored to each child’s developmental stage.
Designed with a screen-free reading approach, Kidodo protects children’s eyesight while encouraging deeper focus and imagination. Its soft, friendly form and interchangeable “outfit system” strengthen emotional attachment, allowing the product’s appearance to evolve alongside the child.
By combining advanced AI capabilities with thoughtful, child-centered design, Kidodo redefines intelligent education as warm, safe, and emotionally supportive—positioning AI not as a replacement, but as a trusted learning companion throughout early childhood.
Project Innovation/Need
Innovation | Dual Innovation in Technology and Emotion
Kidodo is an AI-powered learning companion for children, built on a real-time multimodal large model. It not only “sees and hears,” but understands semantics, recognizes emotions, and generates personalized responses. The product integrates an end-to-end multimodal AI architecture with children's educational scenarios, achieving a full interactive loop of visual, auditory, and emotional feedback. Its innovative PrivateLoRA technology localizes privacy computing, setting a new standard for “privacy-first” family AI products. Kidodo's AI emotion system includes 100 eye expressions and 48 emotional states, giving it human-like empathy.
Kidodo's self-developed model delivers:
1.. Low-latency multidimensional recognition: 350ms response time to keep children focused.
2. Real-time recognition: No pre-input needed for knowledge identification.
3. High intelligence: 76B parameters, enabling reasoning abilities.
4. Personalized memory: Learns user preferences and learning progress over time.
Product Differentiation | Design & Experience Innovation
Kidodo shifts from traditional "teaching" models to immersive, child-friendly interactions, with an innovative “wearable system” that lets children change its appearance, fostering emotional connections. Safety features, like food-grade silicone ears and a soft light camera cover, demonstrate our commitment to user experience.
Brand Differentiation | AI with Heart
As AutoArk's first hardware product, Kidodo embodies the “AI for Growth” philosophy, transforming technological innovation into emotional value and becoming a connector for parent-child relationships.
Design Challenge
Design Challenge / Market Need
This project aims to create an intelligent early-learning companion that combines emotional companionship with cognitive stimulation for families with children aged 0–10. Kidodo responds to global challenges in early childhood education, including unequal access to language stimulation, limited parent-child interaction, and the scarcity of high-quality multilingual learning resources.
Across regions, families face similar constraints. In China, many young children are primarily cared for by grandparents, leading to gaps in language and cognitive engagement. In South Korea, the cost of early education and English immersion programs places a heavy financial burden on families. In Japan, long caregiving hours and limited shared family time reduce opportunities for meaningful interaction. In Europe, multilingual exposure is common, yet consistent home-based learning support remains fragmented. In North America, parents value reading-centered education, but shared reading time is declining while tutoring costs continue to rise. Globally, parents wish to remain actively involved in their children’s development but are often constrained by time and energy.
Kidodo is designed in response—as an AI family companion that can accompany, interact, and educate. Powered by a multimodal real-time interaction model integrating vision, speech, and emotion recognition, Kidodo naturally engages with children. It reads any physical book or object aloud, transforming daily moments into learning experiences. With support for 23 languages, a screen-free design, and a soft, emotionally engaging form with interchangeable outfits, Kidodo brings warmth, focus, and trust to AI-powered early education.
Sustainability
Sustainability Considerations
Kidodo was envisioned from the start with long-term companionship and sustainable growth at its core, balancing environmental responsibility, emotional durability, and social value.
Environmental Responsibility
Kidodo’s shell is made from recyclable, non-toxic, and durable materials that meet child safety standards (RoHS/REACH). The fabric cover is detachable, washable, and replaceable, extending product lifespan and reducing waste. Its modular structure allows hardware upgrades and OTA system updates to be performed independently, enabling users to access new functions and model optimizations without replacing the entire device—significantly lowering electronic waste. Packaging is produced from recycled paper with simplified printing processes to reduce carbon footprint.
Sustainable Design
Designed as a growing product, Kidodo is built upon in-depth research into key developmental stages of children aged 0–10. A structured age-based system ensures that each function precisely responds to the cognitive, emotional, and educational needs of each stage. Beyond technical scalability, Kidodo emphasizes emotional longevity. Its interchangeable “outfit system” allows the appearance to evolve with a child’s age and personality, discouraging short-term use and premature disposal. Long-term usability and a repair-friendly design further support sustainable consumption.
Social Benefit
Kidodo addresses educational accessibility and family companionship by offering low-barrier, high-quality intelligent learning support through natural voice and visual interaction. It helps reduce parental pressure while supporting children’s language development and emotional awareness. Advocating “gentle technology,” Kidodo positions AI as a bridge for parent-child communication rather than a replacement, reducing screen dependency and reinforcing technology’s warm, human-centered role in childhood development.
Product Design - Education - International
This award celebrates creative and innovative design for either a component or overall product. Consideration given to aspects that relate to human usage, aesthetics, selection of components and materials, and the resolution of assembly, manufacturing and the overall function.
More Details