Blog Page

In the “Top Gun, HAL 9000, and Jobs of the Future” podcast (September 15, 2022), WBC director Robert J. Marks discusses a theme from his new book Non-Computable You with talk show host Michael Medved: Can drones should replace pilots in warfare? Dr. Marks, a professor of computer engineering at Baylor University, is also the author of The Case for Killer Robots:
A partial transcript, notes, and Additional Resources follow.
Michael Medved: Why shouldn’t we be able to replace all those hotshot pilots, like the ones being trained in the movie Top Gun: Maverick — one of the most successful movies ever made, by the way, in terms of its box office receipts? That’s showing pilots doing death-defying, astonishing jobs and guiding their airplanes. Well, can’t you do it better using artificial intelligence and drones? …
When you think about people, not just in movies, but in real life, people in Ukraine right now who are risking their lives; wouldn’t it be just a tremendous blessing for us to replace all the human service people, who sacrifice their lives, with machines that could probably do every bit as effective a job?
Robert J. Marks: Involving humans in this was very 20^th-century. The original Top Gun came out in the 20th century. But they were just trying to repeat that same old scenario.
Note: Top Gun: Maverick came out in 1986. A successor film, of the same name, also featuring Tom Cruise, thirty years later, came out under the same name in 2022.
Robert J. Marks: But it turns out we have the technology now, with drones. We have drones that can be controlled by pilots remotely. So they can see where they’re going.
This was, for example, the technology that we used to take out General Soleimani a while back. We used a drone. We fired a hellfire missile and just took him out without any danger of human life.
Robert J. Marks: That same thing could have happened on the mission of the Maverick: the Top Gun sequel. Bunch of advantages. One is the g-force. One of the big things in the movie was that the g-forces would black you out:
Robert J. Marks: Well, that’s not a problem with drones that don’t have any people on them. You can really accelerate. And as long as you don’t destroy the computer, your AI is still there.
Another aspect was that they thought that they needed to fly below where they could do GPS and things like that. Well, with modern military, you don’t need GPS. Some of the military radar drones have maps on the inside. They take pictures of the terrain on the outside. They compare the pictures of the terrain with the maps, and they can go ahead and they can navigate without any GPS.
So, there a number of different aspects which lead one to believe that the technology used in Maverick: Top Gun (2022) was really ill-advised. It should have been done with the 21^st century technology, with artificial intelligence, drones, and even autonomous killing weapons. It would’ve been more realistic, but I don’t think it would’ve been as fun of a plot.
Michael Medved: And you certainly couldn’t have featured Tom Cruise playing a drone. And that’s a big part of it. Just think of the movie posters.
Note: So far, in the real world, AI has been performing better than pilots. From Dr. Marks: After Thursday’s dogfight, it’s clear: DARPA gets AI right. In the dogfight Thursday between AI and a pilot, AI won. But what does that mean? By posing relevant questions, DARPA’s overall AI strategy accurately embraces both the capabilities and limitations of AI. (August 23, 2020)
Michael Medved: If you do it with all machines, basically… all the machine movies, usually, end up showing some of the limitations of machines.
But one of the things that I was thinking about is, in warfare, we used to be so worried. You talk about the 20^th century; about a mistake that leads to a nuclear apocalypse. And wouldn’t further use of AI and machines, to actually not only fight on the battlefield but also to guide strategy and to help direct the command structure, wouldn’t using AI in that reduce the risk of mistake?
Robert J. Marks: Well, one of the things you have to be careful about, I think, is that AI is very brittle. And one of the things that has to be applied in the command field is to be able to react to situations which you’ve never seen before. If you don’t have those in your training data, and you are exposed to a total different scenario that you’ve never seen before, you won’t know how to react if you’re AI. It’s going to take human intelligence to do that.
And as far as the danger, absolutely. One of the things that needs to be done in the design of artificial intelligence is to apply AI ethics. In other words, make sure that the AI that you design does exactly what it is intended to do, but does no more. And that can be done through extensive testing of domain expertise.
You’re never going to get a hundred percent, but I think it’s probably going to be like the legal system. You have to make sure that the operation is above and beyond any reasonable doubt, in the same level of certainty that you have in convicting a murderer or something. So, I think we can borrow from there and apply these standards to some of this artificial intelligence.
Here’s the second part: AI is not taking away our jobs — because it can’t do them. Robert J. Marks talks with KSCJ talk show host Mark Hahn about HAL 9000 and the opportunities and fundamental limits of AI. One key difference between humans and AI is that humans can deal with new information. AI can only address information for which it is programmed.
You may also wish to read: Top Gun Maverick: Thrilling but outdated by today’s AI. How realistic is the continued use of manned aircraft in light of today’s technology? Humans certainly beat drones when a creative solution is required but if a problem can be described and solved in a textbook, drones make more sense. (Robert J. Marks)
Podcast Transcript Download
Mind Matters features original news and analysis at the intersection of artificial and natural intelligence. Through articles and podcasts, it explores issues, challenges, and controversies relating to human and artificial intelligence from a perspective that values the unique capabilities of human beings. Mind Matters is published by the Walter Bradley Center for Natural and Artificial Intelligence.

source