FAA to Use AI, But Not Most Effectively

WASHINGTON - March 12, 2025 - PRLog -- In response to recent airplane accidents, and the growing number of instances of potentially fatal collisions just narrowly averted, the FAA has announced that it will begin using AI to try to reduce these deadly safety dangers.

But the FAA's AI plan seem to be limited to using such computer systems solely to mine (analyze) data of past near-fatal collisions to identify so-called "hot spots" where close encounters between aircraft occur frequently, and to then take individualized steps at some locations to reduce the risks to passengers.

But taking these corrective steps at a few selected airports is likely to be costly as well as take significant time; and especially with the major shortage of trained airport flight controllers, result in only limited safety improvement in the near future.

But this powerful new computer technology can also be used to provide real-time potential-collision alerts to traffic controllers at all major airports very inexpensively in only a matter of months, argues Professor John Banzhaf, an MIT-educated engineer with two U.S. patents and considerable experience with computers and vehicle safety.

His idea in a nutshell is to use existing AI software to monitor airport radio transmissions, and use it to then warn air traffic controllers of possible potential collisions or near misses; eventually also possibly providing the AI-equipped computer with real-time input from other existing technologies including ground-based radar, digital cameras, and complex target-analytics software already in use and tested in airports such as Miami's.

After all, Banzhaf notes, computers running existing AI programs can do many things no air traffic controller can possibly do, and do it in milliseconds, including:
■ monitor the dozens of different relevant frequencies, including commercial, helicopter, small craft civilian, military, Homeland Security, other nearby airports, etc.
■ in the case at Reagan Airport for example, simultaneously monitor all radar at DCA as well as nearby from Bolling AFB, plus any from the Pentagon, White House, etc.
■ calculate various flight paths and foresee possible crashes in less than a second
and then alert an air traffic controller if there seems to be a problem.

While the FAA is to be congratulated for finally agreeing to use AI to prevent airplane crashes and reduce near misses, it should nor limit the use of this tremendously powerful technology to data mining to identify prior "hot spots," suggests the professor.

Rather, at the very least, it should carefully consider and evaluate the possibility of using AI programs to monitor airplane radio traffic and, along with possible additional inputs, provide in milliseconds real-time warnings of possible impending collisions to air traffic controllers.

http://banzhaf.net/ jbanzhaf3ATgmail.com @profbanzhaf