American tech giants Google and Amazon, along with the Indian Infosys, are increasingly “open-sourcing” their artificial intelligence projects. This approach, in contrast to the more competitive and secretive approach being pursued by many companies China, may help to reduce some risks of A.I. and improve development of the entire field. At Google, the company shared its open source machine learning system TensorFlow in 2015, which has been used in various capacities by companies including Uber, AirBnB, eBay, and more. Amazon has also made its internally-developed courses in machine learning available to developers in order to “consolidate best practices.” IBM has also made its Fabric for Deep Learning software available on GitHub, where it is included with a series of tools to support customization and further development by data scientists, developers, and others. OpenAI, the nonprofit founded by Elon Musk and Sam Altman to form a check to unknown dangers of general AI development also publish papers and open source code, working to support safe A.I. development worldwide. Article author Akshaya Asokan argues that these practices and developments are “a good sign of organic growth in the field because AI enthusiasts believe phasing out black box model will eventually de-escalate the risks associted with it. This will also allow coders to understand the breadth and depth of the technology and use it for the larger good.”

Source: Why Google, Infosys & Amazon Are Betting High On Open-Source AI