Pilots Did Not Stop Automation. They Made It Work for Them.
What professional truck drivers can learn from the pilots before autonomous freight becomes the business model
By Micheal “Cupcake” Cobb
Automation is coming to trucking. Hell, the truth is that automation has already been here for decades. The mistake we keep making is acting like we only have two choices. We can either fight every piece of technology that comes along, or we can roll over and let the technology companies and carriers do whatever they want. Those are not our only choices. We can work with the technology while making damn sure professional drivers have a say in how it is used, who is responsible for it, how safe it has to be, and who gets the money when it makes the operation more productive.
Airline pilots already went through their version of this fight. Autopilot did not make pilots disappear. Modern airliners can handle a tremendous amount of the physical flying through automation, but there are still qualified pilots in the cockpit. Those pilots monitor the systems, make decisions, communicate, deal with weather, recognize when the automation is doing something wrong, and take control when necessary. According to the Bureau of Labor Statistics, the median wage for airline pilots, copilots, and flight engineers was $232,140 in May 2025.[24] They are not making that money because they hold the controls every second. They are making it because they are trained professionals who remain responsible for the aircraft and everybody aboard it. They are making it because they came together. They are making it because they realized decades ago that you either work together, or you do not work at all.
That is the part trucking needs to understand. If a professional driver turns on an automated driving system while traveling on a highway where that system is designed to operate, the driver does not suddenly become worthless. Somebody still has to understand the truck, inspect it, monitor the system, recognize a problem, handle the load, deal with customers and law enforcement, respond to emergencies, and take responsibility when the real world does something the computer did not expect. The work changes, but the responsibility does not disappear. If pilots can use autopilot and still be treated and paid as professionals, why should a commercial motor vehicle operator be treated like a disposable steering wheel holder because the truck can perform some of the driving?
This is not a fairy tale where the pilots saved every aviation job. They did not. Navigators disappeared, and most flight engineer positions disappeared as aircraft and automation changed. Some flight engineers and navigators moved into pilot positions, but other people lost their jobs. The pilots also had major internal fights over how many people should remain in the cockpit. The lesson is not that organization can freeze every job in place forever. The lesson is that organized workers have a much better chance of keeping control over their profession, claiming the new work, protecting their pay, and deciding how the transition happens.
That is where we need to go. We are not going to stop technology from developing, and yelling about driverless trucks will not stop companies from spending billions of dollars trying to build them. We need to make automation work for professional drivers instead of allowing it to be used against them. We need to keep qualified people in the seats wherever the operation still requires them, build professional driver positions around the new systems, demand proof before anybody calls the technology safe, and make sure the productivity gains do not all go to the carrier and the technology company.
Automation did not begin with driverless trucks
Automation goes back much farther than electronics. In 1788, James Watt fitted a centrifugal governor to a Boulton and Watt steam engine. As the engine sped up, the spinning balls moved outward and the mechanism reduced the steam going into the engine. As it slowed down, the mechanism allowed more steam in. The machine used its own speed to correct how it was operating without waiting for a worker to make the adjustment. That made Watt's governor one of the most important early uses of automatic feedback in industry, but it was not the first form of automation. Ancient water clocks used float regulators to control water flow, and people built clocks, automata, mill machinery, and other mechanical devices that performed tasks automatically long before 1788. Watt's governor matters because it put self-correcting control to work in major industrial machinery.[31][32]
People talk about autonomous trucks like automation showed up last Tuesday. It did not. I went through machine trades in high school, and automation was already part of machining long before most people ever heard the term artificial intelligence. Numerical control machines were being demonstrated at MIT in the early 1950s. The operator used coded instructions to control the movement of the cutting tools. Computer numerical control followed, and later conversational systems such as ProtoTRAK made the machines easier and faster to program.[22] The machine could repeat the same cut all day, but it still took a knowledgeable person to select the tooling, set up the part, understand the material, check the measurements, spot a problem, and decide whether the finished product was right. Automation changed what the machinist did. It did not make knowledge, judgment, or responsibility worthless.
Ralph Teetor is a perfect example of how long automated control has been part of driving. Teetor was a blind automotive engineer who got the idea in 1936 while riding with his friend and patent attorney, Harry Lindsey. Lindsey would speed up and slow down depending on who was talking, so Teetor started working on a device that could help hold a selected road speed. He filed his application on August 11, 1948, and received U.S. Patent No. 2,519,859 on August 22, 1950.[23][33] Perfect Circle called the device the Speedostat. Chrysler introduced it on several 1958 models under the name Auto-Pilot, and Cadillac used the name Cruise Control in 1959. That means a vehicle system taking over part of speed control was on American roads long before anybody was talking about artificial intelligence or autonomous freight.
Cruise control took over a repetitive part of driving, but nobody looked at it and decided the driver had become useless. The driver still selected the speed, watched the road, steered, braked, judged traffic, and remained responsible for the vehicle. We accepted that kind of automation because it made a tiring task easier without pretending the machine had taken over the entire job. Automatic transmissions followed the same basic idea. The transmission chooses the gears, but the driver still operates the vehicle. EPA data show how normal that has become: manual transmissions have accounted for less than one percent of new U.S. vehicle production since model year 2021.[34] Millions of people use automatic shifting every day without stopping to call it automation, because once a technology becomes familiar, we quit acting like it is some strange new threat.
The same thing is happening under the hood. Most modern gasoline engines do not depend on a person working a choke or manually adjusting the fuel mixture and spark as conditions change. Sensors feed information to a computer, and the computer keeps adjusting fuel delivery, ignition timing, idle speed, emissions equipment, and other engine functions while the vehicle is moving. Modern heavy duty diesel engines package electronic control modules, high pressure fuel systems, variable geometry turbochargers, engine braking, aftertreatment, predictive cruise control, road speed governing, idle control, and automated transmission features into the powertrain.[35] The exact setup changes from one engine and truck to another, but the basic point does not: computers are already making thousands of small decisions about fuel, air, boost, shifting, braking, and emissions while we drive. The person behind the wheel has been working with automation for years.
We use the same principle all through everyday life. Set a thermostat to 70 degrees and it checks the temperature, turns the furnace or air conditioner on, and shuts it back off when the house reaches the setting. Nobody wants to climb off the couch every five minutes to operate the furnace by hand. A washing machine fills, agitates, drains, and spins through a programmed cycle. Automatic doors sense that somebody is approaching and open. Our phones can read an article aloud while we drive, and an iPhone can speak either selected text or everything on the screen.[36] People ask Siri or Alexa to find something, play music, change a connected television, turn on lights, or run a routine. We do not reject those systems because they perform an action for us. We use them because they save time and effort.
There is a difference between automation, mechanization, and remote control, but all three are part of the same long march of innovation. A typewriter did not decide what letter to write, but one keypress mechanically formed a consistent letter instead of making a person shape it by hand. A basic television remote did not choose the channel, but it let the person control the machine without getting out of the chair. A voice assistant or scheduled routine goes another step by interpreting a command or acting at a set time. Even a barrel valve on a mechanically injected Top Fuel engine shows how mechanical control can perform work that nobody handles as a separate step. As the throttle linkage moves, the barrel valve meters the fuel flow with it, so the engine receives more fuel as the throttle opens. That is not the same kind of self-correcting feedback system as Watt's governor or a thermostat, but it still shows how we build mechanisms to make one action control another automatically.
Computers followed the same path. The first electronic computers could fill rooms or entire buildings. The computer used in the Apollo program was an incredible piece of equipment for its time, but we now carry more general computing power in a laptop or a phone. Technology gets smaller, faster, cheaper, and more capable. That is what technology does. Innovation happens. We are not going to stop that process by refusing to admit it exists, especially when we already use and enjoy automation from the moment the thermostat turns on in the morning until the phone reads something to us at night.
The real question is who the technology serves. Does it make the professional operator safer, more productive, and better paid, or does it become another tool used to cut the driver out of the money? Does it reduce crashes, or does a company simply claim it is safer without releasing enough information for anybody to check? Does it reduce fatigue, or does it force a driver to stare at a system for hours and somehow be ready to take control in a split second? Automation may create real benefits, but those benefits have to be proven, and the people doing the work have to share in them.
How the pilots built power before autopilot took over more of the flying
The Air Line Pilots Association (ALPA) started in 1931 when 24 pilot representatives met at the Morrison Hotel in Chicago. Pilots at that time were dealing with unsafe airplanes, long hours, low pay, weak job protection, and companies that treated them as replaceable. They did not wait until the airline industry was fully developed before organizing. They organized while the rules, the equipment, and the profession were still being built.[1][2]
That timing mattered. The pilots were able to make the case that they were not simply employees moving a set of controls. They were trained professionals whose judgment was part of the safety system. Over time, ALPA built much more than a group that bargained over wages. It built departments for safety, engineering, accident investigation, training, legal work, economics, government affairs, communications, retirement, and contract enforcement.[9] When aviation policy was being discussed, pilots could show up with people who understood the equipment, the regulations, the economics, and the actual job.
That is one of the biggest lessons for us. A labor organization has to bargain over wages, benefits, and working conditions, but a real professional organization also has to understand the safety and technical side of the industry. We cannot walk into a room with engineers, regulators, attorneys, and billion dollar companies carrying nothing but anger. We need drivers, researchers, safety specialists, trainers, accident investigators, economists, and attorneys who can challenge the information, produce our own evidence, and explain what the job actually requires.
Decision 83, the Air Mail Act, and the part AFFTRA is missing
One of the most important pieces of the pilots’ early fight was National Labor Board Decision 83. The decision established an 85 hour monthly flight time limit and created a minimum pay formula that included an hourly base, mileage, and the speed of the aircraft. The reason aircraft speed mattered was simple. A faster airplane could produce more transportation in an hour. The formula made sure pilots received part of the benefit when new technology allowed the company to move more passengers, mail, or freight in less time.[2]
Congress then gave those standards federal force through Section 13 of the Air Mail Act of 1934. Air mail contractors had to follow the National Labor Board’s decisions concerning pilot and mechanic compensation, working conditions, and labor relations. Section 12 also allowed the federal government to establish maximum flying hours and safe operating requirements.[3] Congress did not just look at miles. It connected pay, hours, working conditions, safety, and collective bargaining.
There is an important legal detail here because different sections of the law keep getting mixed together. Section 13 was the labor provision. It made compliance with National Labor Board decisions a condition of receiving, extending, or holding an air mail contract, and it said the law could not be used to restrict collective bargaining. Section 12 gave the Secretary of Commerce authority to set maximum flying hours and safe operating methods. Section 17 gave the Postmaster General authority to cancel a contract when a contractor willfully disregarded the contract or the law. Section 13 created the labor requirement, while Section 17 helped give that requirement teeth. The language about unreasonable or unnecessary mail delays that sometimes gets attributed to Section 13 does not appear in Section 13 of this 1934 act.[3]
Long & Harmon showed both the power and the limits of that protection. The company paid its pilots below the Decision 83 scale, and the pilots turned to ALPA. Federal labor proceedings ordered reinstatement for two pilots and back pay for another, while the Post Office opened proceedings requiring the company to comply or face cancellation of its air mail contract. But this was not a clean ending where the airline simply surrendered and everybody went back to work. Long & Harmon kept operating while arranging to transfer the route to Braniff. Braniff took over the postal obligations but did not take the pilots. The wage rule mattered, and the federal contract gave the government leverage, but the workers still needed enforcement, representation, and protection when the business changed hands.[2]
Congress never completely erased that requirement. The language was carried into later aviation laws and is now found at 49 U.S.C. § 42112. The law still says covered air carriers must maintain pilot and copilot compensation rates, maximum hours, and other working conditions in conformity with Decision 83. It also says pilots can bargain for higher compensation and better working conditions.[26] The exact pay system used by modern airlines has changed, and modern flight and duty limits are now covered by other rules as well, but the basic federal protection remains. Current FAA rules, for example, limit covered flightcrew members to 100 flight hours and 190 flight duty period hours during any 672 consecutive hours.[30]
The Railway Labor Act was also extended to airline employees in 1936. That gave pilots a national process for representation, bargaining, and mediation inside the airline industry.[4] Truck drivers do not automatically fall under that airline labor system, but the history shows how federal labor law and collective bargaining worked together instead of being treated like two separate fights.
This is where AFFTRA’s proposal falls short. AFFTRA is proposing a federal floor of $1 for every dispatched hub mile, loaded or empty, paid to the company driver.[27] Paying for empty and loaded miles is better than paying only loaded miles, but it still measures only the movement of the truck. It does not directly pay the driver for inspections, fueling, waiting for dispatch, loading, unloading, handling paperwork, dealing with a breakdown, sitting through required training, or spending hours at a customer because the customer is not ready.
Mileage measures production. It does not measure all of the labor. That is the hole in a mileage only proposal. A driver can give a company 14 hours of the day and only be paid for the miles that happened during part of that time. If traffic stops, the driver loses. If the shipper takes five hours, the driver loses. If the truck breaks down, the driver loses. If the company sends the driver somewhere and the load is not ready, the driver loses. The driver did not get that time back. The company still controlled where that driver was and what that driver could do, but a mileage only pay system can pretend those hours had no value.
Decision 83 did not make that mistake. It included an hourly foundation and then added mileage and aircraft speed so the pilot shared in the increased production. That is the model we should take into trucking. The driver needs a guaranteed hourly floor for every driving and on duty hour. Mileage, stops, automation responsibility, or other productivity pay can be added on top of the hourly rate, but those payments should never replace payment for the driver’s time.
Congress would have to create a trucking version of that protection. Decision 83 does not automatically apply to truck drivers. The important point is that Congress has already used federal law to establish compensation, hours, and working conditions for workers in a safety sensitive transportation industry. If Congress could do it for pilots and air carriers, Congress can do it for company truck drivers and motor carriers.
Federal law already tracks our time. It should require companies to pay for it.
The federal government already makes drivers account for their time. The hours of service system records driving, on duty not driving, sleeper berth, and off duty time. A property carrying driver generally may drive up to 11 hours after 10 consecutive hours off. The driver cannot keep driving beyond the 14th consecutive hour after coming on duty. A driver working for a carrier that operates every day of the week cannot drive after reaching 70 on duty hours in eight consecutive days and can restart that calculation after at least 34 consecutive hours off.[28]
Most drivers understand the amount of time involved without needing a government study. Four full 70 hour work periods equal 280 hours. The point is not to create another clock. The point is that the government already recognizes driving and on duty time as real enough to regulate for safety. It makes absolutely no sense to say the time counts when the government is enforcing hours of service but does not have to count when the carrier writes the paycheck.
Our company driver proposal fixes that problem. The starting floor is $30 per hour for every driving and on duty hour, with time and a half after 40 hours in a week and $30 for every night the driver is required to remain away from home. Stop pay and other production pay can be added when the job calls for it. Detention does not need to become another complicated side payment if the driver is already being paid by the hour. If the driver is on duty, waiting, inspecting, loading, unloading, handling company business, or required to remain ready to work, the clock keeps paying.
The overnight payment recognizes another part of the job that ordinary hourly work does not cover. A driver may be legally off duty in the sleeper berth, but that driver is still hundreds of miles from home because the company sent the driver there. The driver cannot spend that evening with a spouse, attend a child’s event, work around the house, or sleep in a normal bed. Thirty dollars a night is not an hourly wage for sleeping. It is compensation for the fact that the job requires the employee to live away from home.
This pay floor also gets company drivers closer to an actual livable wage. At $30 per hour with overtime after 40, a 50 hour week pays $1,650 before overnight pay. Across 52 weeks, that is $85,800. A 60 hour week pays $2,100, which equals $109,200 across 52 weeks before overnight pay. Four nights away each week adds another $6,240 per year. Actual yearly income will depend on how many weeks and hours a driver works, but this structure finally pays for the whole workweek instead of hoping the driver can run enough miles to make the numbers work.
The $30 rate is the floor, not the ceiling. A union contract can bargain for more money, better benefits, pensions, stronger overnight pay, paid training, job protection, and a share of the savings created by automation. Many interstate drivers are currently excluded from the Fair Labor Standards Act’s normal overtime requirement because of the federal motor carrier exemption. That means time and a half after 40 would need to be written into a new federal trucking standard or secured through a binding labor agreement.[29]
This part of the proposal is for company drivers. It is not a wage mandate for owner operators. An owner operator owns or leases equipment, pays operating expenses, accepts business risk, and needs a separate freight rate and contract system that covers the cost of running the truck and produces a profit. Company driver wages and owner operator revenue are not the same thing, and trying to force both into one formula would create another mess.
What pilots did after automation entered the cockpit
Pilots did not protect their profession by pretending autopilot was useless. They made the human role part of the rules. In major scheduled airline operations under Part 121, federal regulations still require at least two pilots, with one serving as pilot in command and the other as second in command.[5] The aircraft may use advanced automation, but qualified people still hold legal authority and responsibility.
ALPA continues to support useful aviation technology while opposing plans that would remove a pilot from the flight deck. Its position is that two qualified pilots provide cross checking, shared workload, and protection if one pilot becomes unable to perform the job.[15] That is ALPA’s position, but it shows the line the organization has drawn: use the technology without giving up the trained people responsible for the flight.
FAA guidance says automation is a tool and that responsibility remains with the pilots. Pilots must understand the system’s modes and limits, monitor what it is doing, keep their manual flying skills, recognize when it is doing something unexpected, and intervene when necessary.[6] One pilot may be assigned as the pilot flying and the other as the pilot monitoring. Even when the autopilot is physically controlling the airplane, both pilots still have defined duties.[7]
That is exactly how trucking needs to define the professional operator. The driver’s job should include planning, inspecting, monitoring, checking the system, understanding its limits, communicating, responding to problems, and taking control when needed. A driver sitting in an automated truck cannot be treated as cheap backup labor who gets blamed when the system fails but receives none of the money when it works.
Aviation also learned that people need specific training to work safely with automation. Crew Resource Management was developed around communication, workload, awareness, decision making, and the way people and machines work together.[8] Trucking will need the same kind of serious training. Telling a driver to watch a nearly self driving truck for hours and then grab the wheel with a few seconds of warning is not training.
Safety reporting has to be more than a complaint database
Aviation built systems designed to learn from small problems before those problems became deadly. NASA operates the Aviation Safety Reporting System, which accepts voluntary and confidential reports from pilots, controllers, mechanics, and other aviation workers. Identifying information is removed, reports are studied for patterns, and urgent safety alerts can be issued.[10] The FAA’s Aviation Safety Action Program allows the company, the FAA, and often the employee’s labor organization to review reported safety events under a written agreement.[11]
Aviation also has separate legal protection for workers who report safety violations. AIR21 protects covered aviation employees from retaliation for reporting certain safety problems.[12] Confidential safety reporting, a joint process for reviewing events, and legal protection from retaliation are three different things. Aviation uses all three because workers will not report problems honestly if every report can cost them their job.
Trucking has some pieces of that structure. The Surface Transportation Assistance Act protects covered commercial motor vehicle employees who make safety complaints, accurately report their hours, cooperate with investigations, or refuse certain illegal or dangerously unsafe work.[13] FMCSA also has the National Consumer Complaint Database, but FMCSA describes it as a place to report past safety, service, or discrimination problems for possible investigation.[14] It is not the same as a confidential, deidentified system built to find patterns before another incident happens.
We should push for a Commercial Transportation Safety Reporting System that works more like the aviation system. Drivers, mechanics, dispatchers, roadside inspectors, first responders, and automation workers should be able to report failures and close calls without putting their names into a public company file. The reports should be stripped of identifying information, studied for repeated problems, and used to issue warnings before the same failure hurts or kills somebody.
Serious automated truck incidents should also trigger automatic protection of the evidence. That includes the software version, sensor condition, control commands, video, maintenance and calibration records, dispatch instructions, hours of service records, and the system’s approved operating conditions. A professional driver or labor representative with technical training should have a place in the investigation. The company that owns the system should not be the only party holding the data and explaining what the system supposedly did.
The automated trucking timetable is no longer theoretical
FMCSA was already asking questions about automated commercial vehicles in 2019. The agency recognized that some existing rules did not clearly require a person inside the truck and asked how the rules should change for vehicles using Level 4 and Level 5 automation. It also discussed the operating design domain, which simply means the roads, weather, speeds, traffic conditions, and other limits within which the system is designed to work.[16]
The U.S. Department of Transportation also studied possible workforce effects. Participants in that process expected automation to arrive in stages, not all at once. They pointed out that trucking includes far more work than driving, that local pickup and delivery work may remain even if some highway miles become automated, that maintenance and technical jobs may grow, and that drivers need to help design the training because drivers understand the job.[17] None of that guarantees enough good jobs will appear. It tells us which work needs to be claimed and protected before a company labels it as somebody else’s job.
NHTSA already requires certain manufacturers and operators to report crashes involving automated driving systems and Level 2 driver assistance systems. The agency warns that its public data have limits, including duplicate reports, missing information, different levels of company access to data, and confidential information that is removed.[18] Reporting exists, but the numbers are not complete enough to let a company simply announce that it has driven a certain number of miles and declare the system safe.
Then came the April 30, 2026 announcement involving Aurora and Hirschbach. Aurora announced a nonbinding memorandum of understanding under which Hirschbach intends to purchase 500 trucks powered by the Aurora Driver, with deliveries planned to begin in 2027. Aurora said the proposed arrangement could support 500 million driverless miles through its Driver as a Service model. Hirschbach described a network where autonomous trucks would handle longer routes while traditional drivers concentrated on shorter hauls. Aurora also said experienced Hirschbach drivers helped evaluate the system before driverless operations began on the Dallas to Houston route.[25]
That does not mean 500 trucks have already been bought and delivered. The agreement is nonbinding, and the final deal, number of trucks, and timing could change or fall apart. It still shows where the industry wants to go. The companies are talking about 24 hour truck utilization and a combination of driverless highway freight with human operated shorter routes. Drivers cannot wait until those networks are spread across the country before asking who gets those local jobs, what those jobs pay, and what happens to the people whose long haul work is removed.
Our position should be simple
We should support technology that can be proven to reduce harm and improve the work. We should oppose any deployment that hides safety information, removes professional responsibility, or forces workers to carry the risk while the companies take the money. Professional drivers need a real place in testing, rulemaking, training, incident review, pay standards, and the decision about how automation is used.
That is not surrendering to automation. It is claiming our place in the system before somebody else writes us out of it. We can work with a company that is willing to share data, train and pay drivers, follow clear safety limits, and guarantee that workers share in the productivity gains. If a company refuses those basic protections, then we have a clear reason to oppose the deployment.
Define the jobs before the companies define them out of existence
The first job that needs protected is the qualified driver inside an automated truck. During testing and early operation, that driver must have the authority to disengage the system and refuse to operate when the truck is outside the conditions it was designed for. That driver needs specific training and additional pay because monitoring automation and being ready to take control is a safety responsibility, not an excuse to lower wages.
There will also be response and recovery drivers who travel to disabled automated trucks, deal with road obstructions, cargo problems, enforcement stops, and equipment failures. There will be people inspecting sensors, steering and braking connections, software status, calibration, and event records. Experienced drivers should also be involved in checking routes, terminals, construction zones, weather limits, safe stopping areas, and the points where an automated truck hands the load to a local driver.
First mile and last mile work will still require professional people in many operations because yards, docks, city streets, customers, and job sites are not controlled environments. Training and accident investigation will create other jobs that should belong to experienced drivers. None of these positions should be advertised as a promise that every lost long haul job will be replaced one for one. They probably will not be. That is why we have to fight for staffing numbers, pay, seniority, retraining, placement rights, and protection against companies turning these new jobs into lower paid positions.
Build a real automated commercial vehicle credential
Pilots kept professional authority partly because training and qualification remained important even as the aircraft became more automated. Trucking should develop an Automated Commercial Vehicle Operator endorsement. The training should cover what the system can and cannot do, where it is allowed to operate, sensor and control failures, communications loss, cybersecurity, emergency response, takeover procedures, fatigue, inspection of automated components, and protection of evidence after an incident.
The credential should require simulation, supervised experience, and another evaluation after major software or equipment changes. The driver must have the right and duty to refuse operation outside the system’s approved conditions. Employers using automation to increase production should pay for the training and the time spent in training. The driver should not have to buy another expensive credential just so the company can save money with new equipment.
Make companies prove safety before expanding
The choice should not be one national yes or no decision about autonomous trucks. Deployment should happen in stages, and each stage should have safety requirements that must be met before the operation expands. Early testing should use a trained driver inside the truck, clear operating limits, continuous event recording, the driver’s authority to shut the system down, and an independent review of the results.
Limited hub to hub operation should only occur on routes and in conditions the system has been tested to handle. Qualified local drivers should handle the more complicated first and last portions of the trip. Trained roadside response must be available, and the company must demonstrate that its response system can handle several problems happening at the same time.
The public numbers need to include more than total miles. We need crashes, injuries, near misses, emergency stops, forced takeovers, operations outside approved conditions, law enforcement contacts, work zone performance, recovery events, sensor failures, communications failures, and problems reported by drivers. Those results need to be compared by road type, traffic, speed, weather, and operating conditions. A million easy highway miles in good weather cannot be used to prove what a truck will do in snow, construction, a crowded city, or a customer’s yard.
Put automation inside the labor agreement
Company drivers need enforceable contract language before automation changes their jobs. The carrier should have to provide advance notice and bargain before automated equipment changes staffing, routes, terminals, job classifications, surveillance, discipline, subcontracting, or outsourcing. The agreement should state that driving, safety monitoring, automated truck response, training, inspection, and incident review remain work performed by the bargaining unit or by an agreed replacement classification.
Drivers moving into the new work should keep their seniority. They should receive paid training and wage protection while making the transition. The agreement should include placement rights, severance if no position is available, and a joint technology committee with the ability to stop a deployment that violates the safety agreement. Corporate promises sound wonderful during a press conference and disappear when budgets get tight. Contract language is what can actually be enforced.
Use the right organization for each part of the fight
Public advocacy, professional safety work, representation of owner operators, and collective bargaining are not legally the same thing. Our current 501(c)(4) structure can continue to handle education, public policy, regulatory petitions, research, and advocacy. We can use a separate labor organization to organize company drivers, win recognition, bargain contracts, handle grievances, and enforce the agreement. Federal tax law recognizes labor organizations under section 501(c)(5), while the National Labor Relations Act controls much of the organizing and bargaining process for employee drivers.[20][21]
Owner operators who are legally independent contractors are not employees under the NLRA.[19] They need a separate program dealing with freight rates, contracts, transparency, expenses, insurance, access to freight, and other business issues. We do not have to abandon owner operators to organize company drivers, and we do not have to pretend that one legal structure can solve every problem. We need the right tool for each job.
What we should demand
We should demand a formal driver seat in every federal process dealing with automated commercial vehicles. Professional drivers and labor representatives need to be included in rulemaking, research, exemptions, safety evaluations, and publicly funded demonstrations. The people who actually perform the job should not be invited in after the regulators and companies have already decided everything.
Every automated operation should have written limits covering the roads, speeds, weather, traffic, construction, cargo, communications coverage, and support required for the system to operate. Driver and response worker duties must be clearly defined.
The federal government should require truck specific safety data and independent investigation. Companies receiving grants, tax benefits, special road access, government freight contracts, or other public support should have to provide paid training, wage standards, workforce plans, safety reporting, and protection for affected workers. A company asking taxpayers to help pay for automation should not be allowed to hide the safety information or dump the employment costs back on the same public.
Company drivers should receive the $30 hourly floor, overtime after 40 hours, and overnight compensation. Automation or mileage payments should be added to that floor. The law should also protect against misclassification so a carrier cannot call a tightly controlled safety driver an independent contractor just to avoid wages, unemployment insurance, workers’ compensation, benefits, and collective bargaining.
Drivers displaced by automation need a funded transition program. Automation companies and carriers receiving the savings should help pay for retraining, wage protection, placement, apprenticeships, and benefits. Credentials should belong to the worker and be accepted across employers. A driver should not lose years of training records simply because that driver leaves one company’s private computer system.
A realistic two year road map
During the first 90 days, we should adopt a formal automation position, create an Automation and Human Factors Committee, retain labor and transportation counsel, and prepare a regulatory package for FMCSA and NHTSA. We should also begin collecting information directly from drivers about the work they perform beyond steering, the automation problems they already see, the unpaid time technology creates, and the training they would need for new positions.
During the rest of the first year, we should turn those ideas into documents that can actually be used. That includes model labor agreement language, a proposed confidential reporting system, a training and credential outline, standard definitions for automated operations, an evidence protection request for crashes, and a public scorecard for automated truck testing. Any company demonstration we support should require driver involvement, paid training, public safety measurements, independent review, access to incident information, and a clear way to pause unsafe operation.
During the second year, the goal should be to test the model and use real results to improve it. Any qualifying operation should produce a public report covering safety, reliability, driver workload, workforce effects, training, maintenance, law enforcement and emergency response, customer operations, and the actual economics. Success should not be measured by press releases or the number of meetings held. It should be measured by rules changed, agreements signed, information released, drivers trained, jobs protected, unsafe practices corrected, and money put into workers’ paychecks.
The position that can win
Professional drivers do not need to claim that automation can never be safer than a human driver in any situation. If a system can prove that it reduces crashes under specific conditions, then we should learn from it and use what works. We gain credibility by demanding the evidence instead of making claims that can be disproved later. We can admit that automation may remove distracted driving, fatigue, or other human mistakes from some highway miles while still demanding answers about software, sensors, maintenance, weather, construction, cybersecurity, emergency response, and everything else that can go wrong.
Driving down a clear highway is not the entire freight job. The truck still has to be inspected, loaded, unloaded, maintained, recovered when it breaks, handled at customers, and dealt with when the road, weather, cargo, equipment, or people do something unexpected. Somebody must remain responsible. The fight is over whether that somebody will be a trained and fairly paid professional or a low paid person carrying all the blame without any real authority.
The pilots did not remain successful because they defeated automation. They remained successful because they built a profession around it. They organized early. They tied wages to productivity. They made training and qualifications matter. They built safety reporting and accident investigation systems. They bargained contracts and showed up before regulators and Congress with the knowledge to defend their position.
Trucking needs to do the same thing. We are not going to stop automation, but we do not have to hand our profession over to it. We can work with the technology, make it prove itself, keep professional drivers involved, claim the new jobs, and demand our share of the money it creates.
The message is simple: we are not against automation. We are against using automation to remove professional responsibility, hide safety information, and take all the gains away from the people doing the work.
Sources
[1] Air Line Pilots Association, “Our History”. ALPA states that it was founded in 1931 and now represents more than 80,000 pilots.
[2] George E. Hopkins, Flying the Line, Volume I: The First Half Century of the Air Line Pilots Association, ALPA, 1982.
[3] Air Mail Act of 1934, 48 Stat. 933, especially Sections 12 and 13.
[4] National Mediation Board, agency overview and Railway Labor Act mission; see also 45 U.S.C. § 181.
[5] Electronic Code of Federal Regulations, 14 C.F.R. § 121.385, Composition of flight crew.
[6] Federal Aviation Administration, Advisory Circular 120-123, Flightpath Management, November 21, 2022.
[7] Federal Aviation Administration, Advisory Circular 120-71B, Standard Operating Procedures and Pilot Monitoring Duties for Flight Deck Crewmembers, January 10, 2017.
[8] Federal Aviation Administration, Advisory Circular 120-51E, Crew Resource Management Training, January 22, 2004.
[9] Air Line Pilots Association, “Departments” and “Leadership”.
[10] NASA Aviation Safety Reporting System, “Program Briefing”.
[11] Federal Aviation Administration, “Aviation Safety Action Program”.
[14] Federal Motor Carrier Safety Administration, “Report Safety Violations”.
[15] Air Line Pilots Association, “Two Pilots on the Flight Deck”.
[16] Federal Motor Carrier Safety Administration, “Safe Integration of Automated Driving Systems Equipped Commercial Motor Vehicles,” 84 Fed. Reg. 24449, May 28, 2019.
[17] U.S. Department of Transportation, Driving Automation Systems in Long Haul Trucking and Bus Transit: Preliminary Analysis of Potential Workforce Impacts, January 15, 2021, and the related stakeholder proceedings.
[18] National Highway Traffic Safety Administration, “Standing General Order on Crash Reporting”.
[19] 29 U.S.C. § 152, defining “employee” and excluding individuals with independent contractor status.
[20] National Labor Relations Board, “Your Right to Form a Union”.
[21] Internal Revenue Service, “Labor and agricultural organizations”, explaining Section 501(c)(5).
[22] Massachusetts Institute of Technology, MIT Science Reporter: Automatically Programmed Tools, describing MIT’s 1952 numerical control demonstration and the development of APT.
[23] National Inventors Hall of Fame, “Ralph Teetor”, describing Teetor’s cruise control work and 1948 patent filing.
[24] U.S. Bureau of Labor Statistics, “Airline and Commercial Pilots”, reporting May 2025 median wages and pilot responsibilities.
[25] Aurora Innovation, “Leading Carrier Selects Aurora to Scale Autonomous Fleet to 500 Trucks”, April 30, 2026. The announcement describes a nonbinding memorandum of understanding, and its projected scale and timing are not guaranteed.
[26] 49 U.S.C. § 42112, Labor requirements of air carriers, especially subsections (b)(1) and (d).
[27] McCown Co./AFFTRA, “Our Platform”, describing its proposed $1 minimum company driver pay for each dispatched hub mile, loaded or empty, and stating that the proposal is not current federal law.
[28] Electronic Code of Federal Regulations, 49 C.F.R. § 395.2, Definitions, and 49 C.F.R. § 395.3, Maximum driving time for property carrying vehicles; see also FMCSA’s Summary of Hours of Service Regulations.
[29] 29 U.S.C. § 213(b)(1), the federal motor carrier exemption from the Fair Labor Standards Act’s ordinary overtime requirement for employees within the Secretary of Transportation’s hours of service jurisdiction.
[30] Electronic Code of Federal Regulations, 14 C.F.R. § 117.23, Cumulative limitations.
[31] Science Museum Group, “Rotative Steam Engine by Boulton and Watt, 1788”, identifying the 1788 “Lap” engine as the first surviving Watt beam engine fitted with the centrifugal governor.
[32] Otto Mayr, The Origins of Feedback Control, MIT Press, 1970, a history of automatic feedback mechanisms before and during the Industrial Revolution.
[33] U.S. Patent No. 2,519,859, “Speed control device for resisting operation of the accelerator”, showing Ralph R. Teetor’s August 11, 1948 filing date and August 22, 1950 grant date.
[34] U.S. Environmental Protection Agency, The 2025 EPA Automotive Trends Report: Fuel Economy and Technology Since 1975, reporting that manual transmissions have represented less than one percent of new vehicle production since model year 2021 and documenting the spread of modern engine and transmission technologies.
[35] Cummins Inc., Truck Maintenance and Operation: X15, June 2025, identifying the X15’s ECM, XPI fuel system, variable geometry turbocharger, engine brake, automated manual transmission integration, predictive cruise control, road speed governor, idle control, and other electronic features.
[36] Apple Support, “Hear iPhone speak the screen, selected text, and typing feedback”, documenting Speak Screen, Speak Selection, and Siri activation of spoken content.
Comments