wildcat.english

translated from: Wildcat no. 116, Summer 2026

Interview with John, Ben and Joe from the New York Times Tech Guild

»generally more exhausted and unhappy due to the use of generative AI«

John: I joined the NYT during the pandemic lockdown in the spring of 2020. I work on the mobile apps team. We work on the infrastructure for the various news, games, cooking apps, and so on. These apps are an important part of the NYT’s business.

In the summer of 2020, workers organized a walkout because the NYT published an article during the George Floyd protests calling for a military response against the demonstrators. I participated in the walkout and got involved in the underground campaign to form a union soon after.

Joe: I develop design and engineering standards for the web and mobile apps. I was active on the union’s organizing committee, and then I was on the picket line.

Ben: I’m a software engineer and have been with the NYT for 14 years. My team is responsible for building and maintaining data pipelines. I’ve been involved in the union effort since 2019, when I was an organizer. I'm currently chair of our union's AI committee. I’m also on the AI advisory committee of the Communication Workers of America (CWA).

Why did you, as tech workers, form your own union at the NYT in 2022?

John: The NYT has about 6,000 employees. A little less than half of them are union members, with 1,500 in the NYT Guild. These are mainly journalists and editors, but also security personnel and production workers in the print shop, as well as advertising and sales staff. Our union is called the NYT Tech Guild, which represents about 700 employees: software engineering, product and project management, design, analytics. We are the largest union of tech workers in the U.S. to have gone on strike and won a collective bargaining agreement. We’ve had to learn to exercise our own leverage more than we might have if, from the beginning, we had been a minority group within a union made up mainly of journalists.

Joe: With the rise of digital offerings, the NYT began creating hundreds of new job categories—new jobs that aren’t covered by the union.

John: That’s why the union was relatively weak from the 1990s through the early 2010s. Unionized jobs became a minority within the overall workforce.

Does your NYT Tech Guild collective bargaining agreement also apply to other NYT tech workers?

John: The collective bargaining agreement applies only to full-time workers who live in the U.S. and do not hold supervisor positions, but we’ve seen that it sets standards and policies that affect more workers. Also, workers have started taking actions that go beyond the formal boundaries of any union. For example, at the team or working group level, workers have started to make our own mutual agreements regarding how we will or will not use AI. Contractors have gotten involved because there’s an awareness that we’re all part of a shared work environment.

How many colleagues do you work directly with on the job?

Ben: When we started organizing, the majority of us worked in the office. You worked with a lot of people there, and there were plenty of meetings. You could also talk to people from many different departments in the cafeteria. With COVID and remote work, we had to find new ways to do things. There were a lot of virtual meetings; we had to do a lot of individual organizing, create contact lists… It wasn’t an easy time. And when the offices reopened, many people continued to work remotely, just like me.

Joe: Our current collective bargaining agreement has been in effect since January 2025. It stipulates that employees must be present in the office at least two days a week, but there are exceptions. About 40 percent of our union works fully remote.

What about AI applications? Is there a »before« and »after« ChatGPT for you?

Ben: In 2025, the NYT gave us official access to Microsoft Co-Pilot and Cursor, a generative AI coding tool that supports various models. Then, in March 2026, they introduced two tools: Claude Code and DX. DX is a tool that monitors your productivity against opaque industry benchmarks. And one of those benchmarks is AI usage.

I haven’t experienced a single improvement—we aren’t producing anything faster or reaching our goals any quicker. People are working harder to get the same amount of work done—pushing code, handling pull requests… The workload is intensifying without resulting in the delivery of any additional, useful business features. However, a lot more code is being produced than before. I believe we’re delivering lower quality—even though we aren’t even faster.

It’s different with the young new hires. The NYT is currently hiring a lot of junior and mid-level people who have never programmed without generative AI! My team consists of people who can’t program without Claude. But our work is a craft, and it’s very hard to understand that when you go all-in on AI.

The media is currently reporting the opposite—namely, that Claude and others are leading to fewer and fewer juniors being hired. Can you explain that?

Ben: At least at the NYT they're hiring new people at the lower end of the pay scale. So they have a hard time attracting senior programmers. There’s a lot of turnover. I suspect they see Claude as a silver bullet, helping less experienced programmers perform at a higher level. But it doesn't work. We have to deal with all the extra nonsense that comes with generated code. I have to devote much more of my time to reviewing others' work, so adding in the AI doesn't deliver efficiency. It makes the work of existing senior and staff engineers harder.

People are talking about »AI fatigue.« Have you witnessed AI-driven layoffs at the NYT?

Ben: People are generally more exhausted and unhappy due to the use of generative AI.

John: In at least one case, the company directly admitted that someone was laid off due to AI automation. But we’re not seeing mass layoffs caused by AI. Right now, it’s very hard to tell whether someone is being laid off because they’re being replaced by AI or because the company is cutting costs related to AI.

You organized a strike in November 2024? What were the reasons?

John: Basically, it was about wage equity, work-life balance, remote work, job security—and just cause, because the company tried for as long as possible to prevent our first collective bargaining agreement. We had to do something that would force them to the negotiating table. At first, we had a sort of top-down approach to organizing for the one-and-a-half-year collective bargaining campaign; we sent organizers to talk people into joining our actions. That approach eventually hit its limits. We were able to overcome those limits by organizing directly within our work teams. That really shook people up; a willingness to take action emerged. Ultimately, we really wanted to go on strike—but for many, it wasn’t clear how a tech worker strike would unfold or how it could generate power.

How did you organize during the strike?

John: In New York, there were picket lines in front of the offices in Midtown. The atmosphere there during the strike was very charged, emotional, and full of solidarity. For remote workers, it was a bigger challenge, because there had never been a remote picket line before. People found a way to do something together. For example, they created a website for strike games. Then, in-person meetings began to take place among remote strikers. There’s a high concentration of NYT employees in the Bay Area on the West Coast and in North Carolina, where they met up during the strike. Some making plans to come to New York...

Ben: During the strike, I was on the remote picket line every day. We met in video chats during large Zoom calls. That gave us a sense of being involved. We worked our way through it, opened virtual breakout rooms, and had interesting discussions in smaller groups. Over time—and especially with the live broadcast from the office picket line and the speeches—people began to feel it was important to be there, especially alongside their colleagues on the call.

Was there any repression?

John: One management tactic was moralizing: they claimed that the NYT’s mission—democracy—was being jeopardized by the strike. Another was offering financial inducements: they offered money to anyone who secretly wanted to break the strike. They also hired 300 temporary workers at short notice to break the strike. The most important lesson is that none of this worked. The attempts to undermine our strike were unsuccessful. The participation of our colleagues in the strike corresponded almost exactly to the 95 percent strike vote.

How did the strike end?

John: After the first week, many had seen that the strike was gaining momentum. A strong majority of workers wanted to continue it. But the union leadership had conceived the strike as only a short-term measure to exert pressure around the time of the election. [The strike ran from November 4 to 11, 2024; the U.S. presidential election was on November 5.]

They insisted that it had to end after that. And so it ended without a collective bargaining agreement. Soon after, we ratified a deal with some improvements, but without drastic progress compared to management’s last offer before the strike. I won't downplay what we achieved—we fought hard for the protections and safeguards in the contract. But during the strike, we saw that we have the potential to achieve even more. We’re taking these insights with us as we prepare for the 2028 collective bargaining negotiations.

Ben: Every member should know what we’re fighting for and get involved—so that we can take the strongest possible action. I hope that in the next collective bargaining round, we’ll be able to employ significantly more aggressive strike tactics.