Google Is Now Helping the Pentagon Build AI for Drones

Since quite recently there has been uproar within the premises of Google Inc as its employees protest the tech giant’s collaboration with the US Department of Defense. It has been recently made known to the employees via a companywide email that Google will be developing artificial intelligence to help the DoD in analyzing the countless footages which are collected on a daily basis by the multiple drones deployed by the department.

The major reason behind the uproar by the employees is their stance on collaborating in creation a of product which might be used for surveillance as well as in military combat and thus causing loss of life. The project with which Google has allegedly agreed to work had been name Maven, the main objective of which is to analyze the footage collected everyday from a number of drones and looking through them for important information while ignoring the mundane. This collaboration has also raised concerns about the ethics of machine learning and the development being made in the field. Majority of the employees are concerned how technology would now make it easier for the military to kill people.

However, official statement from Google says that these concerns are baseless as the technology which will be developed will not be used anywhere near military combat. Project Maven or the AWCFT (Algorithmic Warfare Cross-Functional Team) is aimed at speeding up the process of involving machine learning and big data, specifically for the cause of surveillance efficiency. It is approximated that in 2017 alone, nearly 7.4 billion USD were spent on this technology. (the Wall Street Journal reported.)

The Pentagon has a record number of aerial drones under its authority. They all send up enormous amount of data everyday which has to be processed and analyzed to find any anomalies or threats. But it seems that the deluge of information being sent is getting too big to be handled by human analysts who find it impossible to keep up with the data being sent their way. So far the defense department has been able to enhance the sensors in the drones so as to gather more relevant data during the drone’s flight but no advancement has been made yet in the analyzer tools at the disposal of the Pentagon.

Maven was launched with a view to be able to use computers for advanced analysis and be able to sift through footage to narrow down vehicle number and 38 more such separate categories. Maven has also been able to use computers technology to keep a tab on individuals as they come and go from the vision range of the drones. This won’t be the first time that artificial intelligence will be used by the government authorities.

Both military applications and law enforcement have previously integrated artificial intelligence in their working. But here lies the problem. Experts have already detected that AI is not neutral and has random bias of its own while treating the individuals in its system. In 2016, such an algorithm was proven to show racial bias.

However Google is saying that its involvement is limited to providing project Maven with TensorFlow APIs. These would helps analysts in integrating machine learning to identify different objects in the footage. A spokesperson for the company acknowledged the implications of using AI for the purpose of surveillance and perhaps war and has said that the company would ensure that the technology is guided by certain terms and conditions of use. It is not yet known if Google is the only private technical company which has been used for this purpose.