• Skip to primary navigation
  • Skip to main content
General Aviation News

General Aviation News

Because flying is cool

  • Pictures of the Day
    • Submit Picture of the Day
  • Stories
    • News
    • Features
    • Opinion
    • Products
    • NTSB Accidents
    • ASRS Reports
  • Comments
  • Classifieds
    • Place Classified Ad
  • Events
  • Digital Archives
  • Subscribe
  • Show Search
Hide Search

Will you have an AI co-pilot one day?

By General Aviation News Staff · October 7, 2023 ·

Imagine you’re flying your airplane, but you aren’t the only one in control.

Your co-pilot is a computer, which also has its “hands” on the controls.

You and your co-pilot are on the lookout for different things.

If you are both paying attention to the same thing, you — as the human — are in control. But if you get distracted or miss something, the computer takes over.

That’s the idea behind Air-Guardian, a system developed by researchers at the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL).

As pilots grapple with an onslaught of information from multiple monitors, especially during critical moments, Air-Guardian acts as a proactive copilot, researchers said, explaining the partnership between human and machine is rooted in understanding attention.

With Air-Guardian, a computer program can track where a human pilot is looking using eye-tracking technology, so it can better understand what the pilot is focusing on. This helps the computer make better decisions that are in line with what the pilot is doing or intending to do. (Image by Alex Shipps/MIT CSAIL via Midjourney)

But how does it determine attention?

It begins with eye-tracking, according to the researchers. In addition, it relies on the neural system and something called “saliency maps,” which pinpoint where attention is directed.

The maps serve as visual guides highlighting key regions within an image, aiding in grasping and deciphering the behavior of intricate algorithms, researchers explained.

Air-Guardian identifies early signs of potential risks through these attention markers, instead of only intervening during safety breaches like traditional autopilot systems, according to the researchers.

“An exciting feature of our method is its differentiability,” says Lianhao Yin, an MIT CSAIL researcher and lead author on a new paper about Air-Guardian. “Our cooperative layer and the entire end-to-end process can be trained. We specifically chose the causal continuous-depth neural network model because of its dynamic features in mapping attention.”

“Another unique aspect is adaptability,” he continued. “The Air-Guardian system isn’t rigid — it can be adjusted based on the situation’s demands, ensuring a balanced partnership between human and machine.”

In field tests, both the pilot and the system made decisions based on the same raw images when navigating to the target waypoint.

Air-Guardian’s success was gauged based on the cumulative rewards earned during flight and a shorter path to the waypoint. The guardian reduced the risk level of the flights and increased the success rate of navigating to target points, the researchers discovered.

For future adoption, there’s a need to refine the human-machine interface, the researchers said. Feedback suggests an indicator, like a bar, might be more intuitive to signify when the guardian system takes control.

Air-Guardian heralds a new age of safer skies, offering a reliable safety net for those moments when human attention wavers, according to MIT officials.

“The Air-Guardian system highlights the synergy between human expertise and machine learning, furthering the objective of using machine learning to augment pilots in challenging scenarios and reduce operational errors,” says Daniela Rus, a professor of electrical engineering and computer science at MIT, director of CSAIL, and senior author on the paper.

“One of the most interesting outcomes of using a visual attention metric in this work is the potential for allowing earlier interventions and greater interpretability by human pilots,” says Stephanie Gil, assistant professor of computer science at Harvard University, who was not involved in the work. “This showcases a great example of how AI can be used to work with a human, lowering the barrier for achieving trust by using natural communication mechanisms between the human and the AI system.”

The research was partially funded by the U.S. Air Force (USAF) Research Laboratory, the USAF Artificial Intelligence Accelerator, the Boeing Co., and the Office of Naval Research.

For more information: MIT.edu

Reader Interactions

Share this story

  • Share on Twitter Share on Twitter
  • Share on Facebook Share on Facebook
  • Share on LinkedIn Share on LinkedIn
  • Share on Reddit Share on Reddit
  • Share via Email Share via Email

Join 110,000 readers each month and get the latest news and entertainment from the world of general aviation direct to your inbox, daily. Sign up here.

Curious to know what fellow pilots think on random stories on the General Aviation News website? Click on our Recent Comments page to find out. Read our Comment Policy here.

Comments

  1. Brian Russell says

    October 9, 2023 at 8:23 am

    Pretty soon the joy of flying GA airplanes will be gone completely. Seems to me that this world that we’ve created will be taken over by AI or will simply implode. Sure glad I’m well past the middle of my life and hopefully won’t be around to experience it.

    • JimH in CA says

      October 9, 2023 at 5:25 pm

      The good thing about GA, is that we DON’T have to add AI or an autopilot, or glass panels.
      I like my ‘steam gauge’ aircraft….trim wheel, johnson bar flaps, etc.

      So, Garmin identified a ‘pitch trim runaway’ issue…wow !! That sounds like the old mcas 737.!

  2. JimH in CA says

    October 7, 2023 at 5:48 pm

    If it can prevent a gear up landing and force a pilot to land and refuel vs exhausting the fuel, then it’s a good add to the ‘right seat’.
    Otherwise, ‘keep your damn hands off the controls !! ‘.

    Heck, the existing autopilots will disconnect if they lose the airspeed data, but won’t/can’t look at the attitude, VSI and engine instruments to tell that the aircraft is flying ‘ ok’.
    ie; Air France pitot tubes iced over…a/p disconnects.
    737 Max bad aoa data….a/p pitches over, vs the attitude and vsi , airspeed !

  3. Phil says

    October 7, 2023 at 11:11 am

    You can nickname the system Otto.

    • Bob Littlefield says

      October 9, 2023 at 7:44 am

      No other name will do!

© 2025 Flyer Media, Inc. All rights reserved. Privacy Policy.

  • About
  • Advertise
  • Comment Policy
  • Contact Us
  • Privacy Policy
  • Writer’s Guidelines
  • Photographer’s Guidelines