Google Gemini will now be used on the US military’s classified data

The updated agreement between Google and the US Department of Defense allows the company’s Gemini artificial intelligence models to be used for any purpose deemed “legal” by the government. According to the information provided by The Information, this scope is not limited to publicly available data; It also includes studies conducted on classified data.

Previously, Google had collaborated to allow Gemini models to be used on unclassified government data. However, the new regulation expands the scope of the existing agreement, paving the way for Pentagon employees to use these artificial intelligence systems in tasks related to classified data. It is particularly emphasized that the use in question should remain within the “legal” framework; However, this statement has previously created controversy in collaborations between technology companies and defense institutions.

There was a serious wave of objection to this development within the company. According to the information in the news, more than 600 Google employees demanded the rejection of the agreement in a letter addressed to CEO Sundar Pichai. Employees argue that working with classified data, in particular, can increase the risk of misuse of artificial intelligence technologies. Accordingly, it is stated that it would be a safer approach to stay away from such projects.

Security and control discussions in Google’s defense collaborations

In the statement made by Google, it is stated that the company is part of a large group of technology and artificial intelligence providers serving in the field of national security. In addition, the company reiterates that certain limits should be maintained in the use of artificial intelligence and draws attention to the importance of human control, especially in areas such as mass surveillance and autonomous weapon systems. However, the content of the agreement raises questions about how such principles will be implemented in practice.

One of the striking articles of the agreement is that Google is obliged to adapt artificial intelligence security settings and filters in line with government requests. This situation raises the possibility that the ethical and technical limits set by the company itself may be stretched. However, it is also stated that Google does not have any control or veto power over government operations that are considered legal.

In addition to all this, the integration of artificial intelligence technologies with public institutions has accelerated in recent years. Especially in the USA, collaborations with large technology companies in both defense and intelligence fields are becoming increasingly common. Despite this, employee reactions and public debates show that such agreements have not only technical but also ethical and political dimensions.