by Andrew Watts
Maria looked at Karpinsky and they both shook their heads.
Maria said, “No. That’s not possible. I think I read an article in Wired magazine about that. Some engineer claimed to have changed the aircraft’s trajectory. But that’s just not realistic.”
“Why not?”
“Oh, I remember this one. We looked into that,” Karpinsky said. “We think the engineer in the article was able to access that aircraft through the IFE system.”
“IFE?”
Karpinsky sighed the way IT people did when talking to “non-computer” people. “The in-flight entertainment system. He probably used the port in his seat to get into the IFE system. Then, on his computer, he saw some data that he thought was the avionics system data, and he tried to manipulate it.”
“The article I read said that he did manipulate it.”
Karpinsky rolled his eyes. “That’s not possible. The data is routed through a different bus.”
“So then how—”
“Okay, let me put it this way. My daughter has a pretend steering wheel on her car seat. When we go to the grocery store, she turns the wheel and she thinks she’s moving our car, but she isn’t. That’s what happened with this guy. The data bus we use on commercial airliners is a secured, closed system. It is not possible for a passenger to hack into it.”
“And the security breach didn’t affect the Fend 100 that was flying that day?”
“Correct,” Karpinsky said. “We analyzed the electronic data stats for hydraulics, avionics, electrical and mechanical controls, along with their built-in redundant systems and fail-safes. Any interference with the normal signal would be detected and flagged to us. We’re clear to proceed with our test flight.”
“And when is that flight?”
“Four days from now.”
They took a break around five-thirty p.m. Sandwiches and drinks were brought in from a local place down in San Marco.
Flynn asked to speak to one of the test pilots who worked on the program. They brought in a guy named Tim Hutson. Karpinsky stepped out for a meeting.
“So are you going to be flying the Fend 100 during the big test flight in four days?”
“That’s right.”
“And you used to fly for the Air Force?”
“That’s right. C-141s, among others. I was a test pilot before I retired and started working here.”
“Excellent. I saw a bunch of Air Force jets at an air show once. What do they call them? The white ones that fly together?”
“The Thunderbirds?”
“Yeah. That’s it. They were incredible. Wow. Hey, thank you for your service.”
“Thanks.”
Flynn looked at Tim. “Can I ask you something? As a pilot, what do you think of all this automated flight stuff?”
Tim said, “I think it’s going to put guys like me out of business.”
“Really?”
“Sure. I mean, autopilot functions have been around for decades. And they’re getting better and better. Hell, now any kid can buy their own quadcopter drone and control it with a smartphone.”
“So you’re saying that the technology is already there.”
“Yeah, has been for years.”
“So then what gives? Why is everyone making such a big deal out of automating commercial airliners?”
“This is more than just a few safety measures. Your car has cruise control, right?”
“Sure.”
“Okay, well, think of today’s autopilot in planes as being the equivalent of cruise control in cars. The Fend 100 autonomous flight is like those driverless cars you hear about all the tech companies testing out. We’re actually using some of the same technology. There are very sophisticated computers on board that will—potentially—be able to completely replace the pilots.”
“From the tone of your voice, I’m picking up that you’re a bit skeptical?”
Tim said, “Most people want to know that there’s a human being up front, making sure everything goes smoothly. An expert, ready to take charge if anything goes wrong.”
“Do you think pilots are still needed?”
“Right now I do, yeah. I think with any new technology, you’re gonna have bumps. And flying is an unforgiving business. Small mistakes have big consequences. I would want a pilot on board.”
Flynn ate a potato chip. He looked at Maria, who was shaking her head in disagreement. “Maria, what are the arguments for automated commercial airliners?”
“I mean…aside from the massive profits, and it being part of our company’s main growth strategy?”
“Yes. Aside from all of that.”
“Safety. Cost efficiencies. Do you know how many car accidents there are in the United States every year?”
“A lot, right?”
“Yes. Over thirty thousand people die in road crashes every year. Thirty thousand. Think about that number. Not only that, but another two million are injured. Two million people. That’s almost one percent of the entire population. Every stinking year. Compare that to aircraft accidents. The fatalities are in the hundreds.”
“What has this got to do with robot pilots?”
“Her point is going to be that robots make fewer mistakes than people,” Tim said. “She’s saying that bringing autonomous driving and flying to market will save lives.”
Maria nodded. “It will save lives. Lots of them.”
Flynn looked at Tim. “You seem skeptical. You still don’t trust a robot to do your pilot job.”
“That’s because he’s a dinosaur.” The gruff voice of Bradley Karpinsky, who stood in the doorway, wearing his light blue lab coat.
Maria looked at her watch. “Your meeting go okay?”
He plopped down at the head of the conference table. “Yes, thanks. Maria’s right. Computers are smarter than people. It’s not even close. Tim here”—he pointed at the pilot—“is going to get tired. He’s going to drink alcohol. He’s going to forget things. He’s going to have to pee. He’s going to get sick. Distracted. Sometimes, he’ll underperform.”
“Never,” said Tim. “Except maybe the part about getting drunk. But not while flying.”
“A computer—or a robot, if you will—will never get tired, sick, hungry, or have to go to the bathroom,” Bradley said. “It will be ruthlessly efficient. Do you know that even today, if you’re flying on a commercial airliner, the hardest landings are given to the computers? Did you know that?”
“That’s not entirely true,” Tim said.
“What does he mean?” Flynn asked.
“He’s referring to the fact that—depending on the airline and the type of aircraft—if weather minimums get bad enough, the pilots are required to have autopilot fly the landing,” Tim answered.
Flynn stopped chewing. “Are you serious? What do you mean weather minimums—like if it gets too windy?”
Tim gave him a funny look. “No, sorry. I mean ceiling and visibility. If the clouds get too low…the best way to think of it is when it’s really foggy out. That gets the visibility way down. So in that scenario, when the cloud layer is basically zero, and the visibility is zero, then a lot of airlines require that their pilots use an automated approach. In essence, the computers will fly the landing.”
“No shit. Wow. I had no idea.”
“Most people don’t.”
Flynn said, “What if the computers mess it up?”
“The pilots are right there, ready to take the controls. But in reality, with the improved navigational equipment on board and the reliability of the automated flight software, the computers never make a mistake.”
Flynn said, “So is your Fend 100 as simple as that? You guys just hit the autopilot button and the pilot leaves the plane?”
“There are three modes of operation in the Fend automated flight software,” Bradley said. “Type one—normal, pilot-controlled flight. Nothing out of the ordinary. Type two—remote-controlled flight. A pilot from the ground controls the airplane. This is
the way the military controls its drones. Finally, we have type three—this is the really innovative stuff. This is where machine learning and artificial intelligence come into play. And it’s where Fend actually adds to the value chain.”
“Bradley’s never had trouble talking up his own program,” Maria said.
Karpinsky frowned. “Fend’s type three automated flight means that the computers inside the aircraft are doing everything. They handle all the communications to and from ground controllers, tower controllers, and all other air traffic controllers throughout the flight. Much of this is through data exchanges. But there’s an actual voice action-response mechanism that’ll supplement the pilot. And the machines are learning. Getting better every day with the sum learning of the entire system.”
“So how does a computer know what to say like a pilot would?”
“That part is pretty easy. It’s just a bunch of code. A bunch of if-then statements, so that it knows what to do in every scenario. For instance, before it goes into class bravo airspace, it knows that it has to establish two-way communications with the class bravo air traffic controller. So it won’t enter that airspace until it does.”
“What if the radio failed, and it could never establish communication in the first place?”
“Then it would keep going through the if-then statements until it got to the correct course of action. Tim, as a pilot, what would you do?”
Tim said, “It depends—but if all else failed, I could set the transponder to 7600 and keep making my radio calls in the blind. Then basically keep flying the approach and land on the runway and hope they can hear you.”
“And that is what the computers would have the aircraft do,” Bradley said. “It would automatically know who to talk to, what to say, where and how to fly. No humans required. And it would be taking in data from a variety of sources, including the radios. If the air traffic controller called up the aircraft and told it to turn, our computers would follow orders—as long as there wasn’t a more pressing need, like a fuel emergency. And this is just how a human pilot would behave. Right, Tim?”
“That’s the theory.”
Bradley frowned. “Tim’s a purist. He believes that all planes should have people in them.”
Tim said, “Bradley, let me ask you something. Have you ever used the voice-to-text feature on your phone?”
“Yes.”
“And did it transcribe everything perfectly? I doubt it. I once told my mom I was going to get her a birthday cake, but instead my phone told her that I got her birthday crack.”
“I understand your point.”
Tim said, “Same example—but with typed text messages—if the autocomplete feature on your phone messes up and changes what you meant to type, it might give your text message a completely different meaning. Those are two separate technologies. Voice transcription, and text autocomplete. Billions of people around the world use those technologies. Yet everyone knows of stories where they have failed, leading to unintended consequences.”
“This is true,” Bradley said. “Voice recognition is challenging. But it’s constantly improving. And your phone needs to be able to interpret anything. We’re able to cut down a lot of errors because of the standardized terminology in aviation. Pilots and controllers are expected to use precise terms and phrases.”
“My point isn’t that voice-to-text is a weak link here,” Tim said. “It’s that when you replace a complex human worker with a robot, you’re probably going to have a lot of errors that the human would never make. Because humans are critical thinkers, and we can use all our senses and experiences to help make decisions.”
Jake Flynn listened as the pilot and engineer went on arguing their points to each other. He wondered how it must feel to know that computers were about to make your job obsolete.
After a brief break, they came together again so Flynn could finish up his questions to the Fend 100 project team. It was well after business hours now, and Flynn could see that he was wearing out his welcome. “Okay, talk to me about your upcoming test flight.”
“We’re calling it a final approval flight,” Bradley said. “The type certification has already been granted by the FAA, but this is more about the contract. The FAA wants to make sure our first passenger flight goes well, and they’re going to evaluate us while it happens.”
“So if it goes well, what happens?”
“We get the NextGen contract for autonomous flight. You’re familiar with NextGen?”
Flynn shook his hand back and forth. “I’ve read about it. But refresh me.”
Bradley looked at Maria, who smiled politely. She said, “The Next Generation Air Transportation System. NextGen. It’s the new national airspace system that the United States will adopt over the next ten years. The plans will transform the way planes and air traffic controllers navigate and communicate, and the way the industry manages flights in the United States.”
“What are the major changes?” Flynn asked.
Maria looked at her watch. “How much time do you have? How about I give you the high-level version? Today the US air traffic control system primarily uses radar and radio communications. NextGen will change that to use more GPS and data transfers. It allows planes to shorten their routes by going directly to the airports, instead of using inefficient navigational beacons. The data exchanges will reduce how much time it takes to get information back and forth between the planes and the controllers. Think of it as using text messaging instead of voice communication. Texting saves time, right?”
“Right.”
“Before, controllers had to read out long chains of instructions. The pilots would then read it back. That took up a lot of valuable time. Mistakes were sometimes made. The data exchange method is quicker and more precise.”
“So that’s it? That’s NextGen?”
“There’s a lot more to it than just that,” Bradley said. “But those changes are major changes. It might sound simple, but making it happen is a massive amount of work.”
“So how does Fend Aerospace fit into the NextGen plans?”
“Many companies are part of it. I mean, it’s transforming the entire commercial aviation industry,” Bradley said. “The commercial airline industry alone is almost ten percent of the US GDP. And it’s growing fast.”
“Wow. I didn’t realize it was that big.”
“Prior to NextGen, automated flight wasn’t permitted,” Maria said. “There are still a lot of regulatory hurdles, but we’re lobbying hard, and things are looking up. As long as everything goes well during our FAA final approval flight in a few days, we think there’s a very good chance that Fend Aerospace will be the main contractor for automated flight when that part of NextGen comes online.”
“Won’t other aircraft manufacturers just make their own technology?”
“Well, that’s why this is so important to us as a company,” Maria said. “Fend Aerospace isn’t building technology just to be used in its own planes. We’re creating the hardware and software that allows aircraft to be preprogramed, and—if need be—controlled remotely. All aircraft in the NextGen Automated Flight Program would use this technology.”
Karpinsky said, “Our strategy is to own the entire autonomous flight software platform. All the major aircraft manufacturers and airline companies will become our customers. Think of the Fend 100 as our iPod. But we’re also creating iTunes. So if the FAA decides to use the Fend 100 platform, everyone will buy from us.”
Maria said, “It would be the standardized brain for every aircraft that wants to fly using automated flight in the NextGen aviation system—so all commercial flights, as well as logistics flights. Like FedEx. There are a lot of those, too.”
“So what’s the problem?” Flynn asked.
“Competition. Right now Fend has a leg up. We earned the bid for the US program. But European and Asian markets are modernizing their airways as well. Ideally, we would form the global solution. But that’s probably not going to happen. We wo
uld probably create a licensing agreement with the leading European and Asian companies to share the tech. But first, we have to prove that it can work in the US. Everything culminates in our FAA final approval flight in a few days.”
“I didn’t realize this flight was so important.”
“Oh, yes. If this flight goes well, we’ll advance to the next tranche of funding. And our fifteen-year contract with the government will go into effect. That’s everything for Fend Aerospace.”
“Let me ask you a question. What would happen if all your technologies were given to your competition? Say, through a cyberattack?”
“That would be catastrophic for the company,” Maria said. “But we’ve made sure that can’t happen. After the cyberattack last month, we’ve really upped our IT security standards. The data for the Fend 100 is in the Fort Knox of IT security. And it isn’t accessible through the Internet.”
Flynn nodded. “Good to know.”
12
Max and Renee spent the afternoon doing research and waiting for their phone call with MI-6. After two hours of Max watching Renee type, she sent him out to get some food.
Max returned and cooked her the best frittata she had ever eaten.
“My God, you’re a good cook. Where did you learn that?”
“Europe. But the secret to this was the ingredients. Roast tomatoes and onions from the local farmers’ market. I have to admit this is one of the better ones I’ve made.”
She smacked her lips, finishing every bit of the dish.
Max and Renee sat in the living room of the beach house talking and working until six p.m. He had just gotten off the phone with the local fuel truck owner. The man was filling up his Cirrus now.
“It’s time.” Time to call the female MI-6 contact.
She nodded, sliding over on the couch. “Here. Sit.”
Renee had connected her computer to Max’s phone. It would run a series of programs designed to gather more information about who they were talking to.