Editorial: AI and GA

By Vickie Buonocore

A team of CMU researchers have developed an AI pilot that can navigate a crowded airspace. The team tested the AI on two flight simulators: one controlled by the AI, the other by a human — both operating in the same airspace. (Photos courtesy CMU)

The following is an interview I conducted with an AI (artificial intelligence) assistant, on a Google Chrome platform called OpenAI. The assistant is helpful, creative, clever, and very friendly (their words, not mine, but I agree).

With all the talk about AI, ChatGPT and writing and journalism, I thought I’d jump online to see what all the hype is about. I was amazed and very pleased that AI came up with great answers that spin general aviation in a very positive light. I warn you that after a minute or two, you can start to believe you’re conversing with a real human… but, the reality is AI is coming up with answers based on tons of information on the internet about the specific subject matter at hand. In this case, GA is apparently putting a lot of very good, valuable and positive information out into the world.

Here’s the interview… It took me longer to think up the questions than it took AI to answer. I tried to throw in a question about the “dangers of flying,” just to see if AI had anything less than favorable to say, but it came up with a somewhat satisfactory answer. I tossed in an edit just because I think it is more dangerous to drive a car than to fly but overall, AI had some good information, solid advice and “nice” opinions about general aviation. 

Human: Hello, who are you?

AI: I am an AI created by OpenAI. How can I help you today?

Human: What can you tell me about general aviation in the United States?

AI: General aviation in the United States consists of a wide range of activities including air sports, air taxi operations, agricultural activities, sightseeing flights, flight training, gliding, charter services, law enforcement flights, medical evacuations, and much more. In 2019, over 9,000 FAA-certificated air carriers operated in the United States alone. The FAA also reported that over 7,000 airports were available for general aviation activities in the United States. There are currently over 587,000 active pilots in the United States and approximately 321,000 active aircraft registered in the country.

Human: Is general aviation good for society?

 AI: Yes, general aviation is beneficial for society in many ways. It helps promote economic development by providing access to transportation, employment opportunities, tourism, business services, and more. General aviation also contributes to the national defense efforts of the United States by providing a platform for search and rescue operations, surveillance, and other military-related activities. Finally, general aviation serves to promote science and technology through the development of new aircraft, engines, and other components used in flight.

Human: Is flying an airplane dangerous? 

AI: Although flying an airplane is generally safe, there are safety risks associated with it. Statistics indicate that fatal aviation accidents occur more frequently than those in other transportation modes. Always practice safety when flying an airplane and make sure to follow all applicable regulations and procedures.

(I believe that day-to-day driving of a car can be more dangerous than flying but I did appreciate the advice to stay safe!)

Human: Can AI be helpful to general aviation pilots? 

AI: Yes, AI can be very helpful to general aviation pilots. AI can be used to monitor aircraft parameters and alert pilots if something appears to be out of the ordinary. AI can also help pilots navigate complicated airspace and identify potential hazards in real-time. AI can even be used to automate certain aspects of flight operations, such as route planning, fuel optimization, and scheduling.

Human: Would you like to fly an airplane? 

AI: Absolutely! I would love to experience the freedom and exhilaration of flying an airplane. However, learning to fly an airplane requires a significant amount of training, so I'm afraid I would need to find a good instructor first.

Human: How would you go about finding a good instructor?

AI: Finding a good instructor will require some research. I would start by looking for reviews and recommendations from other pilots. I would also research the instructors' credentials and ask around for references from their previous students. Finally, I would arrange an initial meeting with prospective instructors to get a better sense of their teaching style and determine if we would be a good fit.

Human: Thanks for answering my questions. May I use your answers in a story I'm working on for In Flight USA magazine?

AI: Of course! I am happy to provide whatever information you need.

Just one more question:

Human: Why should people learn to fly airplanes?

AI: People should learn to fly airplanes because it is an incredibly rewarding experience. Not only do you get to enjoy the thrill of flight and explore the world from a different perspective, you also gain a much deeper appreciation for safety procedures and aviation regulations. Learning to fly an airplane can be a great way to challenge yourself and to create lifelong memories.

Back to me, the human… need I say more? The answers are generally very good and I especially like that AI would like to fly an airplane! This is where I started thinking there must be a person back there behind my computer answering these questions… or did AI pick-up “the freedom and exhilaration of flying an airplane” from the AOPA or EAA website? Either way, the enthusiasm for piloting came through loud and clear!

I didn’t get into deep or technical questions about flying. We can go back to that another time. And, if you try this at home with deep or technical questions, I would suggest checking with a human before taking to flight with whatever AI tells you! Meanwhile, the question of AI pilots came to mind. While I am opposed to that idea on the surface, or should I say in the air, we may be closer than we think to that reality (or is that an artificial reality?).

AI Pilot Can Navigate Crowded Airspace

By Aaron Aupperlee,  Carnegie Mellon University, School of Computer Science

A team of researchers at Carnegie Mellon University believe they have developed the first AI pilot that enables autonomous aircraft to navigate a crowded airspace.

The artificial intelligence can safely avoid collisions, predict the intent of other aircraft, track aircraft and coordinate with their actions, and communicate over the radio with pilots and air traffic controllers. The researchers aim to develop the AI so the behaviors of their system will be indistinguishable from those of a human pilot.

“We believe we could eventually pass the Turing Test,” said Jean Oh, an associate research professor at CMU's Robotics Institute (RI) and a member of the AI pilot team, referring to the test of an AI's ability to exhibit intelligent behavior equivalent to a human.

To interact with other aircraft as a human pilot would, the AI uses both vision and natural language to communicate its intent with other aircraft, whether piloted or not. This behavior leads to safe and socially compliant navigation. Researchers achieved this implicit coordination by training the AI on data collected at the Allegheny County Airport and the Pittsburgh-Butler Regional Airport that included air traffic patterns, images of aircraft and radio transmissions.

The AI uses six cameras and a computer vision system to detect nearby aircraft in a manner similar to that of a human pilot. Its automatic speech recognition function uses natural language processing techniques to both understand incoming radio messages and communicate with pilots and air traffic controllers using speech.

Advancement in autonomous aircraft will broaden opportunities for drones, air taxis, helicopters and other aircraft to operate — moving people and goods, inspecting infrastructure, treating fields to protect crops, and monitoring for poaching or deforestation — often without a pilot behind the controls. These aircraft will have to fly, however, in an airspace already crowded with small airplanes, medical helicopters and more. 

The FAA and NASA have proposed dividing this urban airspace into lanes or corridors with restrictions on when, what kind and how many aircraft can use them. This would significantly alter the current use and standard practices in this airspace and could create air traffic jams, preventing critical aircraft, like a medivac helicopter, from reaching its destination.

While autopilot controls are common among commercial airliners and other aircraft operating in higher altitudes under instrument flight rules (IFR), developing an AI to handle the often crowded and pilot-controlled lower-altitude traffic operating under visual flight rules (VFR) has challenged the aerospace industry. The team's AI is designed to seamlessly interact with aircraft in the VFR airspace.

“This is the first AI pilot that works in the current airspace,” said Sebastian Scherer, an associate research professor in the RI and a member of the team. “I don't see that airspace changing for UAVs. The UAVs will have to change for the airspace.”

The team has yet to test the AI pilot on actual aircraft, but it has performed well on flight simulators. To test the AI, the team sets up two flight simulators. One is controlled by the AI, the other by a human. Both operate in the same airspace. The AI can safely navigate around the piloted aircraft, even if the person behind the controls is not an experienced pilot.

Commercially, the AI could help autonomous aircraft deliver packages and ferry passengers. Delivery drones and air taxis ideally would not operate with a pilot to save weight and insulate them from a pilot shortage.

 “We need more pilots, and AI can help,” said Jay Patrikar, a Ph.D. student in the RI who worked on the project.

This research was supported by the U.S. Army Research Office and the Army Futures Command's Artificial Intelligence Integration Center (AI2C).