Google Inc. is the world’s biggest internet search engine and the company has long touted its software for building digital “cognitive services” on the web.
But the software that powers Google’s “CSG” service — a cloud-based “cognition engine” that runs on the Google cloud — has long been seen as a threat to privacy, privacy-conscious consumers and the open web.
In its latest round of research, Google’s research arm found that the CSG “predicts” users of its software to “view and view images of objects that have been made by computers.”
The CSG is used to build “categories” that help users “think about what’s in front of their eyes,” according to the Google Research Center.
“We think this is a problem that we’re working to fix,” Google said in a blog post on Tuesday.
“This is a product of the times and it is important to note that this is not something we can control.”
The company’s research is a part of a broader effort to tackle the security concerns of its “cloud” software, which includes Google Drive, Gmail and Chrome.
The CSAs are built on Google’s open source cloud platform and allow users to search and download images from the company’s cloud.
But some privacy advocates and researchers have criticized the CSGs for allowing data-collection from outside of the company.
Google has said it does not use the CSGS to “cook data” but instead “saves images and other content.”
The new research is part of Google’s efforts to convince consumers to embrace the cloud and to make their own choices about how they use the internet.
“It’s about getting the cloud to be a place where the content we consume is free from government snooping,” Google VP of product management Ben Golub told Business Insider in an interview.
“And we’re very confident that we can deliver that.”
The Google CSG does not “read” the content or store it, Golub said.
It is a collection of hundreds of millions of images and videos from Google’s cloud, which is “not a huge data center,” Golub added.
The researchers used Google’s own data on the CSGA to build a model of users’ reactions to the CSGF.
The models they used to create the CSGB models included a lot of data that was from outside Google’s data center.
“You might see the CSGC as a place to capture information about people who are watching videos, but we’re not actually capturing information about what people are actually watching,” Golbud said.
“If you’re watching something that you can’t watch because it’s not on Google Drive or you’re not logged into Google, we can’t see that.”
Golub noted that the images in the CSGG models were not necessarily from inside Google’s network.
The data is collected on devices that run Google’s web browser, but it is not stored on servers.
Google said the CSRG models could have a “potential privacy impact,” but did not specify how significant the impact might be.
The research was funded by the Google AI Laboratory, the Google Innovation Lab and the Google Brain Research Group.