On December 2, Microsoft donated the first batch of audio content from artificial intelligence speech synthesis to the “mind library” small program developed by Beijing Hongdandan visual impairment cultural service center, which can be accessed for free from now on. These sound contents include a series of classic works by Lu Xun, Lao She, Xiao Hong, Zhu Ziqing, etc. < / P > < p > these contents are based on speech studio and artificial intelligence speech production platform of azure cloud cognitive service. The selected sound sources include customized voice authorized by artist Zhou Xun and Hong Dan visually impaired announcer Dong Lina, as well as the voice of several Microsoft intelligent speech synthesis platforms. After the application of the latest technology, it can synthesize an audio book within hundreds of milliseconds as soon as possible, and it can be generated continuously for 7 * 24 hours. Zeng Yu, founder of Zhou Xun AI voice red Dan public welfare project, said that Microsoft respects intellectual property rights and protects information security, and is a reliable partner. The AI voice synthesized by Microsoft voice technology is very natural, humanized and full of emotion, and the AI voice of Zhou Xun finally presented is persuasive. < p > < p > in 2014, Microsoft and red Dan cooperated to build a “Library of minds” on azure to help the visually impaired listen to audio books from mobile devices at any time. In 2019, Microsoft uses neural network speech synthesis technology to create more natural artificial intelligence sounds such as “Xiaoxiao” and “Yunye”, as well as a platform that can convert text into speech in batches, thus reducing the cost of creating audio content. On April 23, 2020, the world reading day was held by Microsoft and jointly sponsored by Zhou Xun and Hongdan. < / P > < p > the audio book production project is part of the AI for accessibility project of Microsoft’s social responsibility project AI for good. The plan is expected to invest US $25 million in the five years from 2018 to 2022 to develop artificial intelligence technology that can enhance human capabilities, so that people with disabilities around the world can better enter the workplace, integrate into modern life and enhance interpersonal communication. < p > < p > Microsoft’s text to speech technology based on deep neural network has been released in more than 50 regional languages. The current voice model can almost reach the level of live recording, and can provide multi role, emotional and multi style, which is suitable for the platform of rich scenes, native or customized voice. The report shows that the number of app store purchases soared in the first half of this year due to the impact of covid-19