Many large companies do not allow their employees to use ChatGPT. The reason is that OpenAI’s use of the data we input into this AI is distrusted. Although its creators ensure that privacy is guaranteed, the truth is that there is no complete guarantee that our data will be safe from other eyes.

It is not surprising that Microsoft has created an artificial intelligence for the CIA disconnected from the Internet. According to a news report from Bloomberg, this AI is based on Chat GPT 4. Let us remember that the company founded by Bill Gates provides important material resources to Open AI

It is the first time that Chat GPT has been made to work disconnected from the Internet. The United States spy agencies thus ensure that your data is safe. Well, they use their own, closed data network, independent of the Internet. But creating this disconnected AI has not been easy.

Let us remember that the intelligence of Chat GPT is based precisely on the large deployment of computers in the Open AI cloud. Partly driven by Nvidia’s famous H100 cards. To match the computing power that Chat GPT normally uses in Internet-connected computing centers, Microsoft has been adapting a supercomputer located in the state of Iowa for months.

It is probably the same one that Microsoft has in that state and that was used in 2020 to pre-train Chat GPT 4. Its location was in the news then due to the large consumption of water that was needed to cool this powerful machine. By the way, the location in this state is because the temperatures are colder than other locations in the United States.

About 10,000 people are expected to use data from this spy version of Chat GPT. Whose name, by the way, is not known. Its operation is expected to be very similar to that of the GPTs that we can train with the standard GPT 4 Chat. Although on a much larger scale. Obviously this AI is trained with largely confidential information.

Off-grid artificial intelligence could make a lot of sense for many institutions and companies. And even for any citizen, our information is safeguarded from prying eyes. It goes without saying that the cost of an infrastructure like the one created by Microsoft must be meteoric. And not only to launch this version of Chat GPT for intelligence, but also for its maintenance costs.

Offline data processing is one of the next objectives in the deployment of artificial intelligence. Microsoft itself, despite being one of the major participants in OpenAI, is developing an artificial intelligence with the capacity to run on a phone or tablet: Phi 3 Mini, which has a capacity similar to that of ChatGPT 3.5.

One of the great advantages of artificial intelligence without an Internet connection is that they improve latency. In other words, the response capacity of an offline AI is much faster than that of an online AI. This can be great news for using AI in many jobs.

Nvidia already has software capable of running artificial intelligence on a PC using some models of its graphics cards. Chat With RTX is the name of Nvidia’s AI capable of running on a PC with a graphics card from the RTX 30 or RTX 40 range, which can be obtained from around 300 euros.

The latest releases of Apple’s latest ipads and macs also seem to be related to using AI in the machine itself. Well, both the M3 and the new M4 processors have an architecture designed to run artificial intelligence applications. These new features will surely be revealed at Apple’s developer conference.