Added whistleblower protections for AI workers

On Sept. 29, Gov. Gavin Newsom signed into law Senate Bill 53, the Transparency in Frontier Artificial Intelligence Act (TFAIA). Enhanced whistleblower protection for key employees of companies, called large frontier developers, that train AI models on an exceptional level of computing power and that have over $500 million in annual revenue, is one of several guardrails the new law installs against the catastrophic risks of using and deploying powerful AI systems. 

A risk is “catastrophic” under TFAIA if it is foreseeable that development, storage, use, or deployment of a frontier model will materially contribute to the death of, or serious injury to, more than 50 people or cause more than $1 billion in property loss or damage from a single incident in which the model: (1) provides expert-level assistance in the creation or release of chemical, biological, radiological or nuclear weapons; (2) engages in conduct without meaningful human oversight that is either a cyberattack or that would constitute murder, assault, extortion or theft if committed by a human; or (3) evades the control of the model’s developer or user.

Justification for special whistleblower protections

The effective deployment of sophisticated AI systems may bring great rewards, but also may carry great risks. “‘Foundation’ or ‘frontier’ models are the largest, most powerful AI systems being built today,” wrote a Senate Rules Committee analyst. “A frontier model might be used to cure disease or, conversely, to engineer a new pandemic.” 

The analyst continued: “Whistleblower protections within companies developing AI foundation models can be crucial given the extraordinary scale and impact of the technology. These protections serve as an essential safety valve in an industry where the potential consequences of unethical or dangerous development practices can affect millions, if not billions, of people.”

TFAIA’s enhanced whistleblower protection

TFAIA prohibits covered companies from adopting or enforcing any rule or policy that would prevent key employees from, or that would punish covered employees for, disclosing to the California attorney general or a federal authority information about activities of the employer the employee reasonably believes “pose a specific and substantial danger to the public health or safety, resulting from a catastrophic risk” or information that the employer has violated TFAIA. The enhanced whistleblower provision also will apply to disclosure of such information to those within the company with authority over the employee or to another covered employee who has the authority to address the reported issue.

Covered developers also must establish an internal process through which a covered employee may anonymously disclose such information.

Covered employees

TFAIA’s enhanced whistleblower protections do not cover all employees of frontier developers. Only employees “responsible for assessing, managing, or addressing risk of critical safety incidents,” including harm from materialization of a catastrophic risk and loss of control of the model, will be covered under new Labor Code section 1107. All other employees will be covered by the general workplace whistleblower law, Labor Code section 1102.5.

The Rules Committee analyst explained: “When employees with direct knowledge of risks, harms, or unethical practices cannot safely speak up, dangerous systems may be deployed without proper safeguards or public awareness.”

How TFAIA’s protections are different

Labor Code section 1102.5 protects employees who disclose internally or to enforcement agencies information the employee reasonably believes indicates the employer’s violation of a state or federal statute or a local, state or federal rule or regulation. 

TFAIA’s whistleblower provisions protect more than disclosures of perceived unlawful behavior. TFAIA also protects disclosure of catastrophic risks of harms that may not be unlawful. As one legislative committee analyst put it, “AI development is so new and rapidly developing that an employer’s activity may not violate any existing law or regulation.” Put differently, AI models are developing faster than the speed of law. 

In a joint statement opposing the measure as it was headed to final passage, the California Chamber of Commerce and two industry groups criticized this special treatment. “SB 53 rewrites California’s already robust whistleblower protections for just one industry. Creating a special, one-off standard for a single sector not only sets a poor precedent but also risks confusion and inconsistency across industries.”     

Will these new AI rules inhibit innovation, avert disaster, cause confusion? Yes. 

Eaton is a partner with the San Diego law firm of Seltzer Caplan McMahon Vitek where his practice focuses on defending and advising employers. He also is an instructor at the San Diego State University Fowler College of Business where he teaches classes in business ethics and employment law. He may be reached at eaton@scmv.com.