When an Algorithm Recommends Discipline or Firing: What California’s SB 947 Requires

We have already seen rules laid out for how artificial intelligence is used in the hiring process and applicant screening. Now, to complete the circle, the question shifts to what happens after you’re hired: when a software decides your fate at work – flagging you for discipline or recommending you for termination – where exactly is the human in the loop?

Starting July 1, 2027, Senate Bill 947 brings strict new rules to workplaces that use automated decision systems to discipline or fire workers. While hiring algorithms have consistently faced rigorous scrutiny, SB 947 ensures that employees targeted by the algorithm bosses have transparent rights, mandatory human oversight, and a paper trail.

In July, we wrote about the rules that apply when software screens job applicants. Those Civil Rights Council regulations, in effect since October 1, 2025, make it unlawful to use an automated-decision system that discriminates on a basis protected by the Fair Employment and Housing Act, in hiring and in later decisions such as promotion, pay and discharge. SB 947 adds rules that apply when an employer uses a system’s output to discipline or fire an employee, whether or not discrimination is involved.

What the law covers

The law applies to any “automated decision system” (ADS), a computational process built on machine learning, statistical modeling, data analytics or artificial intelligence that generates scores, classifications, or recommendations used to assist or replace human decision-making which materially affects employees. Routine tools like spam email filters, firewalls, antivirus software, calculators, and databases are excluded.

Importantly, the statute covers state and local government employers, such as cities, counties and school districts, and labor contractors, farm labor contractors and foreign labor contractors, with no minimum employee threshold.

A human must corroborate the decision

Under new Labor Code section 1522, an employer may not rely solely on an ADS when making a disciplinary or termination decision. If an algorithm drives the decision, the employer must direct a human to actively corroborate it using supporting and relevant data. This includes looking at supervisory or managerial evaluations, personnel files, work product of employees, peer reviews, and witness interviews.

If the human reviewer finds that the system’s output or the decision it has arrived at is “inaccurate, incomplete, or misleading,” – or if they simply cannot corroborate it – the employer is barred from using it to take the adverse employment action.

The written notice and your data

If an employer primarily relied on an ADS to discipline or fire an employee, they must hand the employee a clear, stand-alone written notice at the same time. The notice must be written in the language the employer uses for routine communications with employees and sent by a simple method such as email. It must state:

• that the employer primarily relied on an ADS to make the decision;
• that a human reviewed the decision and corroborated the system’s output;
• contact information for a person the employee may contact for more information about the decision and about the right to a description of the employee’s own data; and
• that the employer may not retaliate against the employee for exercising these rights.

Your data rights and what you can expect

The law also provides concrete transparency rights. The employee may request, and the employer must provide, a meaningful, objective description of the employee’s own data used by the ADS. When an employer provides this description, any personal information belonging to customers, other employees, or individuals must be anonymized.

Three prohibited uses

Under section 1522, an employer may not use an ADS to:

• violate, or prevent compliance with, federal, state or local labor, occupational health and safety, employment or civil rights laws or regulations;
• infer an employee’s protected status under Government Code section 12940, which lists characteristics such as race, religious creed, disability, sex, gender identity, age and sexual orientation; or
• “predict and take adverse action against a worker for exercising their legal rights.”

Retaliation, enforcement and proof

An employer may not fire, threaten to fire, demote, suspend or otherwise retaliate against an employee for using or trying to use these rights, complaining to the Labor Commissioner, or cooperating in an investigation.

The Labor Commissioner may enforce these rules by investigating violations, ordering temporary relief pending a full investigation or hearing, issuing citations and filing civil actions. Public prosecutors may also enforce them. Each violation carries a $500 civil penalty. In a court action brought by the Commissioner or a public prosecutor, the party bringing the case may seek “appropriate temporary or preliminary injunctive relief, including punitive damages, and reasonable attorney’s fees and costs.”

The law also shifts the burden of proof. In a civil action or administrative proceeding over the human-review rules, once it has been demonstrated that an ADS was used to make the decision, the employer must demonstrate that the employer did not primarily rely upon an ADS or that it complied with sections 1522 and 1524 when making the decision.

Under Labor Code section 98.7, a retaliation complaint to the Labor Commissioner is due within one year after the violation, and that period may be extended for good cause.

Where the Law Draws the Line: Exclusions and Exceptions

Because SB 947 focuses specifically on disciplinary and termination decisions, hiring decisions fall outside its scope. If an applicant believes software unfairly screened them out during the hiring process based on a protected characteristic, their protections come from the Civil Rights Council regulations we discussed previously, which cover employers with five or more employees, public entities of any size, and employment agencies.

The law does not apply to parties covered by a collective bargaining agreement if the agreement explicitly and unambiguously waives these rules, expressly sets out wages and working conditions, and provides its own protection against algorithmic management. The rules also do not apply to automated systems required by, or reasonably necessary to comply with, a federal statute, federal regulation or binding federal contract relating to the development of aircraft for use in the national airspace, or of products or services for national security, military, space or defense purposes.

Questions worth asking

For a decision made on or after July 1, 2027, these questions help show whether the employer followed the law:

• Did your employer rely solely or primarily on a score, classification or recommendation from software to make the decision?
• Were you given a separate written notice when you were told about the decision, and does it give contact information for a person you can ask about the decision and your data?
• Have you asked for the description of your own data used by the system, and kept a copy of your request?
• Has your treatment at work changed since you asked about the system or raised a concern?
• If a union contract covers you, does it expressly waive these rules and meet the other conditions described above?

Keep the notice, any response to your request, and notes of dates and conversations.

When to seek legal guidance

If a decision to discipline or fire you rested on software, or you were treated differently after you questioned it, Java & Jebreil can talk with you about what happened and what to expect at each stage. The firm represents California employees in wrongful termination, discrimination and harassment and workplace retaliation matters.