This would be bargained between management and the union and is not automatic. Maybe it's a threat but not necessarily an issue Amazon would face even with the formation of a union if this is something Amazon really wanted to keep out of a contract.
At the end of the day both the company and the workers/union have to agree on the terms and people may not be willing to strike over this single issue if the rest of the deal was good.
There are plenty of other reasons for Amazon to want to prevent a union from forming including increased pay, benefits and time off but a union's ability to prevent automation may not be huge issue as Amazon is still growing and hiring workers faster than the jobs are being automated.
> Doesn't automation eliminate the need for negotiating with a union?
Not if the union literally forbids you from introducing automation in the first place. This is how the Transit Workers Union prevents the MTA from using technology that's been standard across the industry for decades.
An alternate strategy is to require the employer to pay the union when new technology is introduced. This drives up the cost of technology to the point where it costs more money than simply hiring the manual labor, thereby preserving the union jobs (at the cost of the taxpayers, in the case of the MTA).
> Trade unions, which have closely aligned themselves with Gov. Andrew M. Cuomo and other politicians, have secured deals requiring underground construction work to be staffed by as many as four times more laborers than elsewhere in the world, documents show.
> The critics pointed to several unusual provisions in the labor agreements. One part of Local 147’s deal entitles the union to $450,000 for each tunnel-boring machine used. That is to make up for job losses from “technological advancement,” even though the equipment has been standard for decades.
The TL;DR is that the L (and soon 7) line is automated with something called "communication-based train control". They still have an operator and conductor on every train, even though the train is perfectly capable of driving itself and making station announcements itself.
I think the MTA thought they could kill off the conductor's position when the trains were fully automated, because the operator could take on the conductor's role now that they don't have to actually operate the train. The union did not allow that.
In the end, this is not really the MTA's biggest problem. CBTC was supposed to increase capacity on the L, but it didn't in reality because the MTA doesn't own enough rolling stock to actually offer that capacity, and the terminals on the L don't allow them to turn trains quickly enough to make the signalling system the bottleneck. (The 7 will be better, though, as at least the Hudson Yards terminal has tail tracks that allow the trains to enter the terminal at full speed. The same cannot be said of 8th Ave. on the L, where trains have to crawl into the terminal because the track ends a few feet after the end of the platform.)
This is not strictly true. CBTC was put on the L first since, at the time of procurement in 2004, it was an ideal testbed as a moderate-ridership, fairly long line isolated from the rest of the system.
It was not immediately obvious that the L would become the center of gentrification in the late 2000s and 2010s; in the '70s and '80s it had such low ridership that they were proposing its closure. So they only ordered enough train cars to meet the 2004 ridership demand.
