March 11, 2024: For nearly a decade the American military has been using software incorporating AI (Artificial Intelligence) to select targets for air strikes for a human to approve or modify. In most cases the air force personnel assigned to check what the software selected find there is nothing requiring modification. In 2023 the software-based target selection system averaged was able to quickly identify 85 targets in seven different areas in Iraq and Syria.

This use of enhanced mission planning software has been around for nearly a decade, but the concept has been around for decades until technology and software capabilities reached the point that such systems became reliable and useful.

While some civilians regard AI enhanced target selection as a nightmarish and improbable effort, the reality is that it works for the troops whose lives depend on it. The success in using the IA based software for target selections is being expanded to include other types of intelligence collection. The fact of the matter is that the AI-based system can make accurate decisions more quickly than human analysts. Human operators must constantly monitor this process to watch out for errors made in the process. Some of those errors were caused by faulty software which is ultimately created by human software engineers. There are always humans in the loop, either directly or indirectly.

US Central Command, while responsible for American operations in the Middle East, Central Asia, and some parts of South Asia, has regularly used algorithms to help decide the state of these regions and what actions the United States should take.

Central Command began using Project Maven's computer vision systems in real campaigns after Hamas' surprise attack on Israel during October 2023. Israeli intelligence was criticized for not detecting the Hamas operation. Israeli intelligence did detect the Hamas plans, but Hamas devised a deception that persuaded the Israelis that there was no danger. Successful deception and surprise are one of the most effective military techniques if you can make it work. Hamas did make it work.

Project Maven scrutinizes large quantities of video and still photos of a combat zone, or potential combat zone, looking for patterns that identify or indicate the possibility of combat or violence occurring. The AI bases system is trained to detect such possibilities and provide warnings of attacks or unexpected military or militant movements. The October 2023 Hamas surprise attack was the sort of thing Project Maven could have convincingly predicted. More so than the Israeli intelligence analysis that did not indicate any dangerous actions by Hamas.

Project Maven can monitor trouble spots in the Middle East, Ukraine, and the Far East where North Korea remains an unstable threat. In combat situations Project Maven can determine where targets are and what impact attacking them would have. Project Maven can also examine past combat situations, if there is enough visual and textual data available, and clarify what happened, how it happened and how that contributed to the outcome. This is a form of predicting the past, which is a standard tool for determining if a prediction system will be accurate and used in situations that have not taken place yet.

Currently there are a number of active combat zones in the Middle East and Russia/Ukraine that would be more effectively monitored with a system like Project Maven, while also using Project Maven to see how useful Project Maven would have been in predicting if these situations would occur and how they will develop in the future. Project Maven doesn’t predict the future but can provide likely future developments. Compared to current intelligence prediction techniques, Project Maven would be faster and incorporate actual outcomes to develop a real-time prediction of how these events are likely to evolve. Intelligence organizations already try to do this manually and have done so for a long time. The results have often not been timely enough or accurate enough to be useful to combat commanders or even planning staff. Having an AI give the same warning that humans give also provides intelligence officials a little bit more credibility with their civilian bosses impending trouble.

With Project Maven the combat commanders and their intelligence staff personnel can quickly evaluate actual or potential combat situations and do so in real time, or at least quickly if potential situations are being examined and evaluated for useful predictions. Project Maven will also strive to overcome the old problem of no combat plan surviving contact with the enemy. Warfare has two sides and Project Maven strives to better understand what the enemy is likely to do.

One lesson of military history is that while enemy reactions to your moves are difficult to predict, it would be a major advantage if you could come up with accurate estimates of future reactions by the enemy. Experience has shown that such estimates, at least accurate and useful ones, are nearly impossible to achieve. Project Maven will try to overcome that problem.

Project Maven is already being used but the results are kept secret. If actual or potential enemies know what Project Maven is predicting they will do, the enemy will change their plans. This has always been a problem with intelligence work, but Project Maven does it in more detail and can change predictions in real time to adapt to changes in enemy activity.

Project Maven may prove to be a major new intelligence tool or maybe not because there may be unexpected pitfalls. You’ll never know unless you try and then you may find yourself reminded of the ancient ironic expression, may you live in interesting times.