An A.I. named “ALPHA,” made by a company called Psibernetix, has apparently impressed the U.S. Air Force by repeatedly splashing a (human) fighter pilot in dogfight simulations. I swear I’m not tricking you into reading my movie script pitch here.
Retired U.S. Air Force Colonel Gene “Geno” Lee is a U.S. Fighter Weapons School graduate, an experienced combat pilot, and an instructor who’s apparently trained thousands of other pilots in the American armed services.
He’s shot down his share of targets, simulated and presumably otherwise, but in a series of simulated air combat missions against ALPHA he could not prevail, Lee told the University of Cincinnati Magazine:
I was surprised at how aware and reactive it was. It seemed to be aware of my intentions and reacting instantly to my changes in flight and my missile deployment. It knew how to defeat the shot I was taking. It moved instantly between defensive and offensive actions as needed.
And even just trying to fly against the robot was taxing:
I go home feeling washed out. I’m tired, drained and mentally exhausted. This may be artificial intelligence, but it represents a real challenge.
As for the machine– the only downtime it needs is for refueling and rearmament.
Psibernetix, which let me reiterate appears to be real and not run by a cyberpunk movie’s antagonist, in conjunction with the U.S. Air Force Research Laboratory, is developing adaptive A.I. systems that can effectively think for themselves. Through what’s adorably (ominously?) called “fuzzy” logic, the system basically figures out an immense web of predictions for possible action/reactions, prioritizes them based on plausibility in context, and executes its own plan accordingly.
Put another way; if ALPHA’s flying a fighter jet against a human in another jet, it will consider every possible move the human can make and have a counter ready for all of them. Taken a step further, ALPHA will observe patterns, learn your moves and wreck you before you even get to decide how to proceed.
It’s pretty much live-action statistics. Running a warplane. But only in a simulator, for now.
As Psibernetix founder Nick Ernest explained to Popular Science:
“The secret to ALPHA’s superhuman flying skills is a decision-making system called a genetic fuzzy tree, a subtype of fuzzy logic algorithms. The system approaches complex problems much like a human would.”
It thinks by “breaking the larger task into smaller subtasks, which include high-level tactics, firing, evasion, and defensiveness. By considering only the most relevant variables, it can make complex decisions with extreme speed. As a result, the A.I. can calculate the best maneuvers in a complex, dynamic environment, over 250 times faster than its human opponent can blink.”
The Journal of Defense Management explains that “ALPHA’s current primary objective is to serve as an intelligent hostile force for pilots to train against within the AFSIM [Advanced Framework for SIMulation] environment.”
Sure, as if this super-smart digital fighter pilot is going to allow itself to be constrained to a “simulation” for long. Have the people creating this stuff even seen Avengers: Age Of Ultron? And also Stealth, which was a terrible film that they could definitely be forgiven for not seeing?
Of course ALPHA will get out into the real world. At least if everything goes according to plan, it’s actually supposed to.
“ALPHA is already a deadly opponent to face in these simulated environments,” Ernest said to the University of Cincinnati Magazine, the publication of his alma mater. “The goal is to continue developing ALPHA, to push and extend its capabilities, and perform additional testing against other trained pilots. Fidelity also needs to be increased, which will come in the form of even more realistic aerodynamic and sensor models. ALPHA is fully able to accommodate these additions, and we at Psibernetix look forward to continuing development.”
The benefits of having a robot brain in a real-world fighter jet are obvious. A computer doesn’t get fatigued. It doesn’t get a case of the Mondays. It doesn’t have personal problems to bring to work and it doesn’t have its own set of morals that might interfere with its interpretation of orders.
Unless maybe it gets struck by lightning like in Stealth. But I’m sure the smart people building this stuff will have a workaround before ALPHA gets airborne for real.
“It’s likely that future air combat, requiring reaction times that surpass human capabilities, will integrate AI wingmen...” writes the University of Cincinnati Magazine. “So, AI like ALPHA could simultaneously evade dozens of hostile missiles, take accurate shots at multiple targets, coordinate actions of squad mates, and record and learn from observations of enemy tactics and capabilities.”
For now it’s just the dominant air power in the simulator; knocking off lesser A.I.s and even some of humanity’s finer pilots.
“The A.I. is so fast that it could consider and coordinate the best tactical plan and precise responses, within a dynamic environment, over 250 times faster than ALPHA’s human opponents could blink,” the University of Cincinnati Magazine report details.
The future is now, and it’s equal parts exciting and scary. If you have the time and interest to dive deeper into the specifics and science behind ALPHA’s software, the Journal of Defense Management has a dense report and ResearchGate has more on “fuzzy logic.”