Google had planned to publicly host the large database of chest x-rays released in 2017 by the U.S. National Institutes of Health (NIH) Clinical Center but pulled out just days prior to the project's announcement due to patient privacy concerns, according to an article published online November 15 by the Washington Post.
The Washington Post reported that Google was set to use its cloud services to publicly host the NIH Clinical Center's dataset of 122,000 chest x-rays, which were intended to facilitate development of artificial intelligence (AI) algorithms. In addition to utilizing the database to show how its TensorFlow software could be used to train machine-learning algorithms, Google also planned to make the x-ray data available to outside AI researchers via its cloud, according to the article.
Google employees and the NIH worked together in the summer of 2017 to scrub the personal patient data from the dataset, but on July 19 -- two days prior to Google's planned announcement of the project -- the NIH informed Google that dozens of images had been found to still include personally identifying information. After potential liability concerns were raised by the company's lawyers over possessing and reviewing sensitive health data, Google elected to delete all the images from its servers and withdraw from the project, the Washington Post reported.
After scrubbing all the data, the NIH Clinical Center ultimately released the database in September 2017 using cloud-based storage provider Box.
