Poll: Where Will AI Find Application In GA?


Other AVwebflash Articles


  1. I’ve used the ChatGPT (openai.com/api/) a number of times already for regulation searches. Be sure to ask for the references it’s good at that. Aviation lawyers I would think are the most threatened. Even your local airport properties and security regulation answers are just a couple minutes away now. During meetings you can ask a ChatBot whether or not a new proposal or request is within regulation and it will enlighten everyone in the meeting.

    Other places in Aviation that AI will take over is Dispatching, Scheduling, writing an SMS program, customer service, pay roll, complicated employee benefit packages and taxes. I’ve used the ChatBot for my self-employment tax questions but, just ask it for the specific tax rule. It will give you an interpretation but, I’d rather read it for myself and decide.

  2. Programs like ChatGPT will be used to write articles for aviation magazines, but never for AvWeb, of course.

    • (laugh)!

      Artificial intelligence mainly deals with decision-making from an assemblage of KNOWN SOURCES of data. UNLIKE the AvWeb editors, it sorts out “outliers” and pushes “outside the mainstream” to the rear of choices.

      By contrast, AvWeb editors think “outside the box”–they often take a page from Bobby Kennedy–“Some men see things as they are–I dream of things that never were, and ask WHY NOT?”

      “Artificial Intelligence” will usually make predictions only on precedent. It takes a Paul Bertorelli to question “Why NOT?”

  3. Given that the poll was posed in the future tense, the obvious (but unavailable) answer is “Everywhere”. And since it allows for only one answer, it should have asked, “Where will AI find application in GA first?”

    At this emergent moment in AI, the technology is limited by its input sources to well-understood and well-defined scopes of knowledge. It can’t, for example, yet answer the question, “Is this airplane pretty?”, but it will before too long. Also, it is currently too limited by the quality of its input sources. ChatGPT can be easily manipulated into responding with nonsense and patently false answers.

    Certainly, the first four poll options will all benefit from AI in the near term. If I were sitting in the pointy end of bomb built by the lowest bidders, I’d certainly want every intelligence available offering its input to the finger on the button.

    I don’t think AI will solve all the NP-complete problems in my lifetime, but while ChatGPT is little more than “Eliza” for the early twenty-first century, there is no joystick on the bridge of NCC-1701.

  4. Lots of focus on ChatGPT but there’s other AI-based technology out there, already being applied.

    The most promising that I’m aware of is that being developed by Savvy Aviation to using engine monitor data to spot trends and potential problems and help with diagnosis. An AI-enabled tool that spots failing exhaust valves is already available and in use by many including myself. With the high quality data that can be downloaded from modern engine analyzers, I expect AI to prove valuable in this particular domain.

  5. Commercial aviation and the Locomotive industry use AI (machine learning really) to predict engine failures using rather simple parameters. I often wonder why we cannot do the same thing with GA engines. Savvy Aviation has taken a first step in engine monitoring.