Punching In: Courts Find Little Transparency on Federal Firings

Monday morning musings for workplace watchers

Trump RIF Facts Few, Far Between| California Bill Targets AI Bosses

Ian Kullgren: For all the legal complexities of President Donald Trump’s reorganization of the federal government, administration lawyers still tend to tiptoe around one of the most basic questions in court: Who, exactly, has been fired?

The dynamic was on display in a hearing Thursday where Justice Department attorneys sidestepped how many people have been let go from specific agencies linked to a Trump executive order, frustrating judges at the US Circuit Court of Appeals for the Ninth Circuit as they waded through murky sets of facts.

The exchange showed how closely the Trump administration is still guarding its reduction-in-force plans—and as a result, how much is still unknown to the public.

Two Clinton appointees, Senior Judge William Fletcher and Judge Johnnie Rawlinson, lit into DOJ lawyers for withholding specifics on federal layoffs .

“I don’t think the legal theory is very much in question,” Fletcher said. “What’s in question is what is actually happening.”

Rawlinson pressed further regarding the firings, which are linked to a Trump order.

“What are the agencies doing under the executive order?” she asked. “Now, that’s the question.”

But DOJ attorney Maxwell Baldi remained vague.

“Some agencies are, I believe, proceeding with plans to reduce their staffing levels, and others are not,” Baldi said, pointing judges to an Office of Personnel Management court filing from July that largely restates public information.

In court filings, the Trump administration estimated 40 layoff plans across 17 agencies. But DOJ has since argued that the counting of individual employees who were laid off isn’t so simple because some agencies have revised their plans.

Baldi told the Ninth Circuit panel that the reorganization plans, which unions have asked to review, are only a “snapshot in time” that often changes.

A separate Ninth Circuit panel on Tuesday struggled to make sense of which entity—OPM or individual agencies—fired probationary workers earlier this year.

“There’s quite a bit of factual dispute about whether or not these phone calls were made, or whether there was or was not a direction to fire the employees,” said Judge Morgan Christen, an Obama appointee.

The DOJ attorney responded that individual agencies had fired those workers, meaning the case should be routed through the Federal Labor Relations Authority or the Merit Systems Protection Board.

Christen noted later that the Trump administration has claimed elsewhere that OPM does, in fact, have the authority to fire workers.

Protesters rally outside of the Theodore Roosevelt Federal Building headquarters of the US Office of Personnel Management on February 05, 2025 in Washington, DC.

Photographer: Alex Wong/Getty Images

Chris Marr: State legislators are on the verge of passing a first-of-its-kind California law restricting how businesses use AI to manage, discipline, and fire their workers, but the proposal faces steep tech industry opposition.

The measure (SB 7), which deals with “automated decision systems” in employment, is awaiting review from the state Assembly Appropriations Committee to decide whether the bill reaches a floor vote before the session’s scheduled end on Sept. 12. The Senate passed an earlier version of the bill in June.

The proposal, like a similar bill pending in Massachusetts, is part of labor unions’ and worker advocates’ broader push to establish baseline worker protections related to businesses’ use of artificial intelligence and similar tools, said Annette Bernhardt, director of the technology and work program at the University of California, Berkeley Labor Center.

If the bill becomes law, California employers would be barred from relying primarily on automated systems to make decisions about firing workers, and they’d be required to give workers notice about the types of electronic surveillance and decision-making tools they’re using. Businesses also would have to give workers a chance to appeal for human review of disciplinary decisions made with the help of AI tools.

“None of this is controversial,” Bernhardt said. “Workers should have the right to know which systems their employer is using, and they have the right to expect that it’s a human, not an algorithm, that is making critical decisions impacting their economic future.”

The bill’s sponsors face steep hurdles in overcoming industry opposition, as most states’ efforts to regulate employer AI use in the workplace have fallen short of becoming law. Colorado passed in 2024 the most sweeping state law to date governing AI decision-making tools, but lawmakers there are now considering heavily revising it or delaying implementation at the urging of Gov. Jared Polis (D). The tech industry also is continuing to push for a federal moratorium preempting state laws regulating AI.

The California proposal goes beyond the more widespread—albeit mostly unsuccessful—state legislative proposals aimed at preventing discrimination in AI-assisted hiring decisions. Instead, it targets the many ways employers are using automated systems to manage their employees, track their productivity, and gather data on them to feed into algorithms that make recommendations such as whether they’re likely to be good candidates for promotion.

“More and more businesses are relying on algorithmic management,” said Lorena Gonzalez, president of the California Federation of Labor Unions, AFL-CIO.

The bill’s requirements would help ensure human managers remain involved in key decisions about their employees.

“Bosses should have souls,” she said.

The bill threatens to drive up businesses’ costs and hinder innovation, partly due to the risk of workers suing employers over alleged violations, said Niloy Ray, an attorney with Littler Mendelson PC who advises companies on their use of AI tools.

The proposal also fails to set a reliability standard for the AI tools that companies use—a key but tough-to-solve problem when attempting to govern AI use, he said.

SB 7 is “getting us closer to common sense and feasible regulation,” Ray said. “It’s not there yet though.”

We’re punching out. Daily Labor Report subscribers please check in for updates during the week, and feel free to reach out to us.