Open office with workers at computer workstations

TL;DR: California's AB-1898 cleared the Privacy Committee on March 18 and is headed to Judiciary. If passed, employers would need to give workers 90 days' notice before rolling out any AI surveillance or automated decision-making tools. They'd also have to maintain an annual inventory of every AI system watching you. Penalties: $500 per employee per violation. This could be the strongest workplace AI transparency law in the country.

The Warning You Never Got

Your employer installed productivity tracking software last month. You found out when your manager cited your "mouse movement patterns" in a performance review.

Nobody told you the camera in the break room uses emotion detection. Nobody mentioned the AI screening your Slack messages. Nobody explained that an algorithm decides who gets promoted.

Assembly Member Nick Schultz wants to change that.

His Assembly Bill 1898, introduced on February 12, 2026, would force California employers to warn workers before the robots start watching. Ninety days in advance. In writing. With a signed acknowledgment form before any tool goes live.

What Counts as "Workplace AI"?

The bill casts a wide net. Two categories:

Automated Decision Systems

Any computational process using machine learning or statistical modeling that "issues outputs used to assist or replace human decision-making" about workers. That covers:

  • Resume screening algorithms
  • Performance scoring systems
  • Promotion prediction tools
  • Termination risk assessments
  • Task assignment automation

Workplace Surveillance Tools

Anything collecting worker data through:

  • Video and audio surveillance
  • Geolocation tracking
  • Continuous time-tracking
  • Biometric monitoring
  • Keystroke and mouse tracking

If a human rubber-stamps what the algorithm suggests? Still counts. The "human in the loop" dodge won't work here.

What Employers Have to Do

AB-1898 creates real obligations, not just suggestions:

90-Day Advance Notice

Before deploying any new workplace AI tool, employers must provide written notice to every worker who "would likely be directly or indirectly affected." Union reps get notified too.

The Notice Must Include:

  • Purpose and justification for using the tool
  • Specific employment decisions it affects (hiring, firing, discipline, wages, benefits)
  • What worker data gets collected, how often, and how long it's stored
  • Plain-language explanation of what data goes in, what analysis happens, what comes out
  • The vendor's name and the product name
  • Any quotas the AI enforces and what happens if you miss them
  • Timeline for when the AI might replace your job entirely
  • Summary of any risk assessment conducted

Signed Acknowledgment Required

Employers cannot deploy the tool until every affected worker signs a form confirming they received and understood the notice. No signature, no surveillance.

Annual Inventory

Starting February 1, 2028, employers must maintain and provide workers with a list of every AI tool currently in use. Updated annually.

Penalties With Teeth

This isn't just a "please be nice" request. The bill includes real enforcement:

  • $500 per employee per violation. That adds up fast for large employers
  • Labor Commissioner can bring enforcement actions
  • Workers can file civil lawsuits for damages
  • Punitive damages available in egregious cases
  • Attorney's fees recoverable by workers who win

For a company with 10,000 employees using undisclosed AI tools? That's $5 million in potential fines per violation.

Where the Bill Stands

AB-1898 cleared the Assembly Privacy Committee on March 18, 2026, with a 10-3 vote. The committee recommended it "do pass and re-refer to Judiciary."

If it becomes law:

  • February 1, 2027: Notice requirements take effect for tools already in use
  • February 1, 2028: First annual inventory due to workers

Employer groups are pushing back. The Ogletree law firm warned clients about "administrative feasibility" concerns for large organizations. Translation: it's hard to be transparent when you've been hiding surveillance for years.

What Workers Can Do Now

Ask Your Employer Directly

Email HR: "Does our company use any AI tools for performance monitoring, hiring decisions, or workplace surveillance? If so, please provide documentation." Get it in writing. Even without AB-1898, the question forces them to think about disclosure.

Document What You Know

Screenshot any productivity trackers on your work computer. Note any mentions of "AI scoring" or "automated reviews" in manager communications. This becomes evidence if AB-1898 passes and your employer didn't disclose.

Check Your State

California leads, but other states are watching. New York City already requires bias audits for AI hiring tools. Illinois requires consent for AI video interviews. Your state might have protections you don't know about.

Support the Bill

If you're a California resident, contact your Assembly member. AB-1898 moves to Judiciary next. Worker voices matter in these hearings.

Why This Matters

Workplace AI surveillance exploded during remote work. Employers bought tools that track keystrokes, monitor webcams, analyze email sentiment, and score "productivity" through algorithmic black boxes.

Most workers have no idea these systems exist. They don't know an algorithm decided they're "at risk" for termination. They don't know their bathroom breaks are being timed by machine learning.

AB-1898 doesn't ban any of this. It just says: tell people first. Give them 90 days to understand what's coming. Let them decide if they want to work under algorithmic surveillance.

That's a low bar. But it's a bar many employers aren't clearing.

References

  1. California Legislature - AB-1898 Bill Text
  2. Ogletree - California Workplace AI Bill Would Impose Hefty Penalties
  3. National Law Review - AB-1898 Penalty Analysis
  4. California Lawyers Association - State Legislation Update March 2026
  5. Transparency Coalition - AI Legislative Update