Artificial intelligence (AI) has been a center of attention for the last few decades. A large amount of training data, improved computational abilities and advanced deep learning algorithms recently favor the wider application of AI, including materials research. Traditional trial-and-error methods for material research are low efficiency and time-consuming. AI especially machine learning-based methods could speed up the process by learning the rules of molecule and structure-effect relations that underlie a dataset and building a model to make predictions. There are several stages for the developments of AI based materials innovation: from the data analysis machine, to data-harvest machine and to data creative machine. And this is quite different from the computational chemistry where the computer is little more than a calculator, employing a hard-coded formula provided by a human expert. The development of AI in materials could be the self-learning and problem-solve assistant. 

 As for solid state electrolytes for Li-ion batteries for example, a structure-property relation for electrolytes is not clear enough and there are thousands of potential candidates which could lead to better performance of the resultant battery. Hence, by employing the know-how electrolytes, AI could use the resulting data from a database to train itself and use an AI-designed model to verify the experimentally successful materials, and then even predict, design and synthesis some new promising solid Li-ion conductors. AI is also suitable for addressing research questions in other materials domain, guiding and accelerating the future development for Material Genome Initiative.