Our view is that the U.S. Government will eventually become the largest customer for all three major cloud platform providers (AWS, Azure, GCP). Without a public-private partnership, the U.S. will fall behind China in the race for all things AI.
What are some of the near-term military use cases related to AI? I spoke at a conference in late 2019 about how the military has deployed a combination of machine learning and computer vision (ESRI was the vendor in this case), to quickly identify objects of interest (potential nuclear missile sites), within global satellite images. We are all familiar with the use of unmanned drones used for capturing data (usually satellite images), as well as for carrying out military strikes. Facial recognition technology is another well-known capability that has both military and commercial use cases (Google Photos, Amazon Rekognition). Today we primarily leverage AI to perform discrete tasks that can be performed with much greater speed and accuracy than any human.
Autonomous military operations. Military operations could become Skynet-like if we removed the human element and went partially or fully autonomous. Humans could define broad parameters and allow the machines to predict, plan and execute military operations. For example: “Skynet – define an operation that would inflict maximum casualties against XYZ country with an upper limit of 10 million people while preserving the physical landscape with a mission duration of 10 days or less. All weaponry is at your disposal. Propose three options and wait for further instruction.” (or take action on the operation with the highest probability of success assuming a probability rate of X”). In our example, Skynet may decide that a series of EMPs deployed over a major city in the target country is the best way to achieve the objective. AWS, Azure and Google Cloud are sufficiently massive to power the various intelligent applications required to carry out such an autonomous operation.
The Big Three. It is likely that these three platforms (AWS, Azure, GCP), will be required to have a degree of interoperability to maximize the force multiplier effect the U.S. military would want to achieve.
A series of ethical questions rather than technology related. The questions at our feet are not technology-related. We have the capability. The relevant questions are ethical such as “How much autonomy do we want to provide intelligent systems when it comes to military operations?” WW3 is less likely to be the result of a Dr. Strangelove like rogue military officer and more likely to be the result of a software error. We link to a recent paper published by the National Security Commission on AI (Click HERE). Note that Google, AWS, Oracle and Microsoft each have a seat at the table (Google two seats by my count).