In a move reminiscent of its Android playbook, Alphabet Inc. (GOOG) open sourced TensorFlow, its new machine learning system, this morning. In plain speech, this means that developers, researchers, and university students can use data from the company's cloud to research or develop custom applications for their products. 

In a post making the announcement, the Mountain View-based company stated that it used TensorFlow for everything "from speech recognition in the Google app, to (formulating) Smart Reply in Inbox, to search in Google Photos." The company further stated that it hoped to accelerate artificial intelligence so that "everyone from academic researchers, to engineers, to hobbyists can exchange ideas much more quickly, through working code rather than just research papers." The move makes sense for Google business-wise as well as it could turn into a licensing profit center for the company down the road.

But, the company may face two problems related to this initiative. 

Who Owns the Data? 

The first one relates to data ownership.

More precisely, who owns the final results of the manipulated data?

While open sourcing Amazon Machine Learning earlier this year, Inc. (AMZNsaid it would have read access to all data models created within its ecosystem. In addition, the service does not allow export or import of model data sets.  As Google's service scales and wide and varied data sets and models are created and used, there is a potential for wider misuse (and propagation) of incorrect data patterns. In the absence of clarification from the company, accountability may become a problem.

Closed and Open Ecosystems

The second one is related to competition and ecosystem. Android gained traction because it was working within the limited confines of a mobile ecosystem. Machine learning and artificial intelligence are fairly large ecosystems and span multiple industries and device genres. In that respect, Google faces increased competition from multiple ends. For example, Apple Inc. (AAPL) has snapped up AI companies in recent times. Similarly, Microsoft Corp. (MSFT) announced Azure Machine Learning, a similar initiative, earlier this year using capabilities available in Microsoft products, such as XBox and Bing. International Business Machines Corp. (IBM) also has Watson Analytics, which enables developers to use Watson's powerful engine.

These companies work within closed ecosystems. In a hardware environment, an open operating system ecosystem can cause problems down the road, as Google discovered with bug fixes in Android. Given that deep learning spans multiple industries, the scope and extent of Google's competition and problems could multiply with an open source AI system. 

The Bottom Line  

Google's TensorFlow is a step in the right direction. Hopefully, the company has learned lessons from its Android experience (which has largely been successful) to better manage large open source ecosystems.