Boeing’s Wisk Is Going Full Robot With Its Electric Air Taxi While Competitors Stick With Human Pilots

Dozens of companies are developing electric air taxis designed to take off and land vertically so they can hopscotch passengers across crowded urban areas. All are hoping that someday computers will fly them, eliminating pilots and their salaries and freeing up their seats to carry another paying passenger. Most air taxi developers are betting that safety regulators will be more comfortable approving their cutting-edge aircraft if there’s a pilot in the cockpit, at least to start.

BoeingBA
-controlled Wisk Aero is going in the opposite direction. It’s sticking with an ambitious plan to go autonomous from the beginning, even as it unveiled Monday the design for a larger, four-seat aircraft that theoretically could accommodate a pilot.

Jonathan Lovegren, the head of Wisk’s autonomy efforts, says they can make a better self-flying aircraft if it’s designed for a robot pilot from the start, and that it’s “misleading” for other companies to say it’ll be simple to just pull the human pilot out of their aircraft somewhere down the line. “The reality is there’s a completely different safety analysis and design assurance process,” Lovegren told Forbes. “It’s a really different aircraft.”

Wisk acknowledges that will take it longer to come to market. Competitors Joby Aviation and Archer Aviation are aiming to launch in 2024. Wisk isn’t sharing its target date publicly but the Mountain View, Calif.-based company believes the new aircraft, which it hasn’t begun to flight-test yet, will be carrying passengers before the end of the decade.

Wisk is planning for its air taxi to fly autonomously along pre-planned flight routes under the supervision of an employee in a ground-control station who will oversee up to three aircraft simultaneously. Its sixth-generation air taxis, which it says will be able to fly 90 miles with safety reserves at a cruising speed of roughly 140 mph, will be outfitted with sensors to detect large birds, aircraft or other hazards and will automatically adjust course to avoid them. The human supervisor will be able to override an air taxi’s decision in an emergency or divert it to land, but they won’t have a control stick to fly it manually.

With all the well-publicized struggles TeslaTSLA
and other automakers have had in perfecting self-driving cars, Wisk is adamant that it isn’t using artificial intelligence to create a “black box” version of a human pilot – the Federal Aviation Administration isn’t ready to evaluate software where there isn’t a predictable output for every input. “The reality is there’s not a way to prove beyond a shadow of a doubt to the FAA that it’s going to do the right thing all the time,” Lovegren says of AI-based systems.

Roughly 90% of the tasks on an airliner are already handled by autopilot and other computerized systems. Lovegren says Wisk is simply building on top of that already approved technology to create a rules-based framework to automate the vast majority of the rest.

That includes responses to emergency situations, in which pilots currently consult manuals with checklists that detail how to respond step-by-step.

“Most of it is procedural,” says Lovegren. “It’s very well understood.”

It’s a reasonable approach to automate checklists and “a lot of the button and switch flipping,” says Ella Atkins, an autonomous systems researcher who heads the aerospace and ocean engineering department at Virginia Tech. The problem will be in the rare situations where there’s an emergency that the system can’t match up with a checklist and the supervisor will have to step in, she says.

“A human on the ground that’s managing multiple aircraft will be much slower to react, to understand the situation on this airplane in real time,” Atkins says. And they may not have any better situational awareness than the aircraft’s software since they’ll only see the same data and video feed.

A dilemma for safety regulators is that it might not be fair to expect the average onboard pilot of these new electric vertical takeoff and landing (EVTOL) aircraft to be able handle an emergency any better.

For urban air taxi services to make a profit — and appreciably reduce ground congestion — the companies believe they’ll need to achieve economies of scale by flying hundreds of aircraft at a high tempo in major urban areas. That means they’ll need a lot of pilots. In 2020, consultancy McKinsey estimated that the industry could require 60,000 pilots by 2028 if they rolled out as planned. They’ll need to recruit and train those pilots as airlines grapple with a shortage.

The airlines will be able to pay higher salaries to retain the most experienced talent.

“These EVTOLs, they can’t have the best pilots in the world,” Atkins says. “They’re not going to be that glider instructor who knows what to do when the engines fail.”

That’s one reason Wisk would like to hand the controls over to a robot, and its competitors say their pilots will be supported by highly automated flight control systems.

The bottom line, says Atkins: “It’s not necessarily less safe for Wisk or other companies to automate all of the checklists in a way that is certifiable with the FAA” and have a remote supervisor “than it is to have a [less] experienced pilot onboard.”

How the FAA will assess the safety of the complicated software in air taxis is a point of concern after the failure of the agency to spot the flaws in a flight-control system that contributed to two crashes of Boeing’s 737 MAX.

Observers have warned for years that the agency needs more computer science experts, which it generally has a hard time competing with Silicon Valley to hire.

Wisk’s sixth-generation air taxi is a descendant of development work begun in 2010 by billionaire Larry Page’s Zee.Aero and later a company called Kitty Hawk that Zee.Aero was folded into. The program was spun out as a joint venture with Boeing in 2019; Kitty Hawk last month announced it was shutting down, ending its efforts to bring a smaller autonomous air taxi to market.

Wisk has been engaged in discussions with the FAA for years on how to prove the safety of autonomous aircraft. “They see us as, I think, setting the direction for a lot of this,” says Lovegren.