The New York Times faces union revolt over AI worker tracking

The internal tools meant to make tech work easier just turned into automated surveillance.

The New York Times is facing formal grievances and unfair labor practice charges from its unionized workforce. The Times Tech Guild, representing around 700 engineers, designers, and data analysts, accuses management of violating their collective bargaining agreement by deploying two internal AI tools to monitor performance and discipline workers.

At the same time, the Times Guild, which covers roughly 1,500 newsroom and commercial staffers, filed separate charges over the publisher's flat refusal to share details about its broader AI roadmaps.

From developer experience to digital quotas

At the heart of the dispute are DX, a developer productivity suite, and Glean, an AI-powered enterprise search engine.

According to Ben Harnett, a Times software engineer and chair of the unit's generative AI committee, DX was initially pitched as a high-level way to improve workflow across teams. But over recent months, management began applying individual benchmarks, turning the data into grounds for disciplinary action.

Harnett noted that workers in reviews have had DX statistics thrown in their faces, with managers pointing out things like logging only one pull request in a given week, allegedly 25 percent below industry norms. The union argues this flattens the nuances of complex engineering into crude, misleading quotas that say nothing about actual code quality.

Meanwhile, Glean aggregates everything from internal wikis and GitHub repos to Google Docs and emails so employees can find information fast. But workers say managers can query it to spy on unfinished drafts and individual activity. The union even believes recent written warnings sent to staff were directly formatted and drafted by Glean, despite the tool's known tendency to hallucinate.

Times spokesperson Danielle Rhoades Ha pushed back, stating the company disagrees with the union's characterization and plans to address the claims through normal contractual channels.

Why it matters

We think this clash exposes an uncomfortable truth about generative AI in modern corporate environments: tools sold to staff as helpful assistants are rapidly being repurposed into granular surveillance cameras.

For knowledge workers and software engineers, the immediate takeaway is cautionary. When management starts treating easy-to-scrape AI signals—like token usage, meeting summaries, or pull request counts—as proxies for productivity, it incentivizes gaming the numbers instead of shipping meaningful work. Judging a senior engineer's output by raw Git commits makes as little sense today as counting lines of code did thirty years ago.

On the industry level, this dispute moves the labor fight well beyond copyright and job replacement. While outlets like ProPublica and McClatchy have seen walkouts and byline strikes over AI-generated text, the Times battle targets the legality of automated workplace surveillance under existing union contracts. If the Tech Guild forces management to curtail individual tracking, enterprise SaaS vendors pitching AI efficiency metrics will face far tighter guardrails across unionized workforces.

Management's move looks clever on a CFO spreadsheet, but it carries massive hidden risks. Feeding hallucination-prone AI digests into disciplinary files creates obvious legal vulnerabilities and poisons workplace trust, driving away top technical talent who refuse to work under an automated microscope.