Possible Negatives
All the benefits that AI can deliver to aviation may come with possible negatives that will have to be addressed. One of these uncertain negatives is based on the fact that AI systems are ultimately created and coded by humans, with all the potential consequences associated with this reality.
“AI, although potent, itself suffers from many potential errors that can be caused by stochastic faults (due to the presence of random variables in its calculations), induced by malicious agents, or appear due to human errors,” Vamvoudakis said. “There are also considerations when it comes to implementing AI in aviation. One concern is the level of autonomy granted to AI systems. Striking a balance between automation and human control is crucial to ensure that pilots maintain their skills and decision-making abilities. Additionally, there are questions regarding the reliability of AI systems in critical situations or unforeseen circumstances. Rigorous testing and certification processes will be necessary before widespread adoption of AI-enabled flight technologies.”
Professor Vamvoudakis isn’t alone in voicing concerns about AI. “Many advanced AI algorithms still require significant testing and evaluation for use in critical aspects of aviation, particularly those that affect the safe operation of aircraft,” Briceno said. “There are potential safety risks in the over-adoption of AI technologies
78 Sep/Oct 2023
without understanding the possible consequences when these technologies fail.”
Then there are the “unknown unknowns” as described by former U.S. Defense Secretary Donald Rumsfeld: “the ones we don’t know we don’t know.” The very newness of AI makes it difficult for humans to know how to assess its abilities fully, let alone trust it to take command. And then there is an unexplained phenomenon called “hallucinations,” when AI programs simply get things wrong.
That’s right: for reasons still unknown by AI programmers, sometimes AI programs make things up when answering questions. Take this example from the New York Times May 1, 2023 article titled, “When A.I. Chatbots Hallucinate.” The newspaper asked ChatGTP, “When did the New York Times first report on artificial intelligence?” ChatGTP replied that this occurred on July 10, 1956, in an article titled “Machines Will Be Capable of Learning, Solving Problems, Scientists Predict,” about a conference at Dartmouth College.
Here’s the problem: according to the Times, “The 1956 conference was real. The article was not. ChatGPT simply made it up.” The newspaper’s first article on AI didn’t come out until 1963.
In fairness to ChatGTP, the AI program may have sourced its answer from other web-based sources that were inaccurate but
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84
