SAE Publishes Driver's Ed Plan for Robotaxi Remote Helpers

It never had formal structure. That's changing now. The Automated Vehicle Safety Consortium at SAE International published a best-practice document this week, and it lays out a full curriculum framework for the people who monitor and assist autonomous vehicles from a computer workstation rather than a driver's seat. So the training for SAE robotaxi remote helpers is finally getting the shape it's always lacked. But don't expect a quick fix; the guide is just the starting point for a field that's still defining itself.

The document, titled Best Practice for ADS Remote Assistance Training and Human Factors, is essentially the missing driver's ed manual for an industry that has been improvising since the first robotaxis hit the road. It covers everything. And it's a hefty brief. You need to know the vehicle's sensors, sure. But there's also the trick of keeping someone alert when there's literally nothing to do for hours at a stretch, which is a problem most manuals don't even bother to mention, let alone solve. So it's a start.

What Remote Operators Actually Do

Before diving into the training recommendations, it helps to clarify the role. These are not remote drivers in the traditional sense. There's no internet-connected steering wheel, no sim-rig setup that lets someone take over the car like a giant video game controller. The paper is explicit about this: remote operators monitor deployed vehicles and step in when the car gets stuck on a decision.

Think of it as helping a vehicle choose the correct lane or accept a suggested path, rather than grabbing the wheel. So the operator is more of a supervisor than a driver. It's a fundamentally different job. That shift changes the required skill set, because you're not reacting to every twist and turn but instead monitoring a system that mostly runs itself, stepping in only when the machine asks for guidance.

Rules of the Road, Minus the Road

One of the more interesting recommendations is that operators need a complete understanding of the rules of the road where the vehicle operates, even if that location is on a different continent. That's a tall order. And it means dynamic signage, lane markings, and local traffic laws, all of which could differ wildly from where the operator is physically sitting, so they can't just assume what they see back home applies anywhere else. But the point stands. It's about context, not just maps.

The paper also stresses that operators should understand the autonomous driving system itself. If you're going to spend your shift staring at camera feeds and sensor visualizations, you need to know what created that data and how the system uses it to navigate the real world. That includes the sensor suite, the decision-making process, and anything specific to the company's particular implementation.

"If you're going to spend your shift looking at visualizations and camera feeds being beamed back to HQ, you ought to know what created the data and how it's used to find a path through the real world."

Simulators Without the Pedals

SAE pushes simulators for training. But not the arcade kind. These are screen-based systems, completely devoid of wheels or pedals, and they're built to guide operators through fictional scenarios or replay real-world examples of genuinely tricky situations from actual streets. It's a safe way to expose people to urban driving chaos. No one gets hurt.

Simulations also serve as an assessment tool. They can weed out candidates who aren't cut out for the job, though the paper doesn't specify exactly how to grade performance. It's more of a general recommendation that simulators should be part of both initial training and ongoing evaluation.

The Retraining Gap

Retraining is another area where the paper makes a strong case. In-person driving has a notorious problem with this, and the document suggests the remote side shouldn't repeat that mistake. Operators should go through refresher courses both to keep their skills sharp and to learn about new features, behaviors, or interface elements that get rolled out over time.

It's one of those common-sense recommendations that sounds painfully obvious, yet in the daily grind of fleet operations it gets shrugged off with alarming frequency. So the oversight can have real teeth. Software updates don't just patch bugs; they can fundamentally change how a vehicle accelerates, brakes, or handles a curve, and the person sitting in the driver's seat needs to know exactly what's different before they're ever thrown into a real-world emergency. Don't wait for the surprise. That's the whole problem.

Do You Need a License to Not Drive?

Here's where things get surprising. The paper explicitly notes that remote operators don't necessarily need to hold a valid driver's license. It makes sense when you think about it, since they're not actually driving in the classical sense. The main overlap between their job and traditional driving is knowledge of applicable regulations, which is already covered in the training recommendations.

The document stops short of mandating it. Individual companies can still require a license and a clean driving record if they prefer, so that choice remains firmly in their hands rather than being imposed from above. It's a business decision, not a hard rule. That flexibility lets robotaxi operators staff their remote assistance teams however they see fit, adapting to local conditions or shifting priorities without regulatory pressure. But the door isn't shut on stricter standards. They're simply optional.

The paper also highlights that the workstation environment and workflows present entirely new problems that have nothing to do with traditional driving skills.

Boredom Is the New Danger

Regular cars deal with driver inattention by monitoring jerky steering inputs or droopy eyelids. Remote operators face the opposite problem: cognitive underload. When there are no actions to take, the brain checks out. The paper suggests teaching operators to recognize reduced performance in themselves and giving them a penalty-free way to request a break.

It's a delicate balance. The paper also mentions the need to manage potentially stressful cognitive workload, so it's not like the job is always boring. The workload can swing dramatically depending on the situation, and operators need to be prepared for both extremes.

Some companies also require remote operators to handle communication with people inside and outside the vehicle, as well as first responders. That's a customer service role layered on top of a decision-making role. It's a heavy load. And it's up to each company to determine how to split those responsibilities, so the operator can't just focus on driving but must also juggle public interaction and emergency coordination with equal skill.

Why This Matters Now

The regulatory environment for autonomous vehicles is still slowly coming together. But the industry isn't waiting. Robotaxi services are already operating in several cities, and every one of them has someone behind a screen watching over the fleet, a human safety net that's quietly become the industry's default answer to the gaps in the rules, so it's not exactly hands-off. Until now, there's been no shared standard for how those people should be trained. None at all.

The AVSC's recommendations are mostly common sense. But they're valuable precisely because they're written down and shared across the industry. That's the real power here. Having a document that companies can point to and say "we follow these guidelines" creates a baseline for safety and accountability, and it's a baseline we've sorely needed in a field where vague promises don't cut it. So simple. So effective.

The Transparency Question

The hope is that this starts a larger conversation about remote operator training. Companies like Waymo and Zoox have been tight-lipped about how their remote assistance teams are trained and what their day-to-day work looks like, so this document could pressure them to open up about their processes. That's the real goal. But it's a start.

The curtain may hide good intentions. But intentions don't drive the cars. Actions do, and in this rapidly shifting landscape, the scoreboard is unforgiving, which makes every undisclosed choice a potential liability that can't be waved away with a smile. Now that a formal framework exists, there's no excuse for keeping those details hidden, no room for vague assurances when the rulebook is sitting right there on the table. The industry has a standard to aim for. The public deserves that much. And they have a right to know exactly how the people supervising robotaxis are prepared for the job, from the first training hour to the final override.