By admin 
AI-generated summary reviewed by our newsroom.
Read our AI Policy.
- AB 1883 would ban employers from using AI that collects neural data and analyzes emotions.
- Supporters warned AI surveillance can reveal immigration, medical and sensitive data.
- Opponents said AB 1883 could restrict AI tools used to detect unsafe driving and theft.
A bill aimed at reducing artificial intelligence surveillance in workplaces has led to fierce debate between competing interest groups.
Read more Camilo Doval’s Yankees Tenure Still Leaves Questions About What Went Wrong
Proponents of the legislation say the surveillance violates workers’ right to privacy and can be discriminatory. Opponents say the tools that would be banned by the measure are part of important workplace safety mechanisms.
Assembly Bill 1883 would prohibit employers from using certain AI surveillance technologies that collect neural data and analyze workers’ emotional state.
According to Assemblymember Isaac Bryan, D-Los Angeles, who authored the bill, these technologies can be used to find workers’ immigration status, medical history and other sensitive information.
“Workplace surveillance is not a recent phenomenon, but today’s workplace surveillance tools differ in their scale, their speed and their invasiveness,” Bryan said in a presentation to the Senate Labor, Public Employment and Retirement Committee. “Employers used to have to monitor video feeds to track workers. Now artificial intelligence models and tech advancements allow them to compile massive amounts of data points and analyze them in real time without that human interaction.”
Supporters raise privacy, discrimination concerns
Representatives from the California Federation of Labor Unions, a co-sponsor of the bill, have said in committee hearings that facial recognition tools can employ discriminatory standards through the use of artificial intelligence algorithms.
According to Labor Federation President Lorena Gonzalez, AI facial recognition and emotional monitoring tools are often trained on photographs of every face available on the internet. As a result, these technologies were overwhelmingly trained using images of white people.
“AI is not a person. It doesn’t have the ability to think about nuance,” Gonzalez said. “Quite frankly, it’s offensive, the standard that we use. The darker an individual is, for example, the harder of a time AI has reading their facial expressions. It does have a racial undertone.”
Samantha Gordon is the chief advocacy officer for TechEquity, an advocacy group focused on accountability in the technology industry, which has lobbied in support of the measure. Gordon said safety concerns are often among the arguments used to oppose regulations on AI surveillance.
“We’ve seen lots of examples where (facial recognition technology) is inaccurate, where it’s been weaponized against vulnerable communities,” Gordon said. “Every time that we hear there are safety concerns, we just want to know … what are you seeing happen? What are you seeing that this is absolutely critical for?”
These technological oversights can have negative consequences for workers, according to Gonzalez. In summer 2024, healthcare company Kaiser Permanente rolled out a tool that attempts to analyze nurses’ tone of voice and level of empathy during patient calls. According to reporting by CalMatters, these surveillance systems gave lower scores to nurses who stayed on the phone longer than 15 minutes and often misjudged the emotional context of their conversations.
Gonzalez said she has witnessed firsthand the essential role that triage calls play in patient care. According to Gonzalez, AI analysis of these conversations would reduce nurses’ ability to provide quality care.
“I had a nurse navigator after my doctor told me I had breast cancer. When she first called me, I was sobbing,” Gonzalez said. “By the end of the call, I was laughing and making jokes because she made me feel better. She was laughing too. If you have a computer that doesn’t understand human interaction, that would come across as though her tone was inappropriate for the phone call.”
Read more Will Howard is Steelers Biggest Winner of Training Camp: Week 1
The Labor Federation is among several organizations that have pledged support for AB 1883, including co-sponsor Teamsters California, the California Teachers Association, Utility Workers Union of America, Service Employees International Union California and the California Nurses Association.
Advocates for the bill say further amendments that permit AI surveillance in the workplace would constitute an invasion of privacy. Tracy Rosenberg, advocacy director for data privacy coalition Oakland Privacy, said that such an overreach would present a “profound” capacity for misuse.
“If we give permission in the name of safety for employers to monitor the thoughts and feelings of their employees, then we are essentially saying that it is more important to protect certain profit-making interests than it is for (workers) to have the freedom of their thoughts and feelings, and I think that’s the wrong equation,” Rosenberg said.
Opponents say AI tools “vitally important” in workplace
Opponents to the bill have cited workplace safety as a primary concern. According to California Manufacturers and Technology Association CEO Lance Hastings, the legislation represents “a broad brush against some technologies that can actually make the workplace safer.”
“With the insurance costs climbing the way that they are and the quick movement of commerce in the modern economy, having these tools is vitally important,” Hastings said. “What we don’t want to see is a step backward in safety at our facilities.”
CMTA joined several dozen other groups, including hospital and grocers’ associations, commercial organizations and county officials, as signatories on a letter of opposition to the measure. A July 2 version of the letter penned by the California Chamber of Commerce says the bill overlooks some “legitimate uses” of AI surveillance in the workplace.
Among the listed uses are detecting unsafe driving behavior in vehicles, identifying theft on workplace premises and flagging “inappropriate behavior” in customer service interactions.
According to John Myers, senior vice president of communications and external affairs for CalChamber, the bill would interfere with the implementation of workplace violence protection plans. Many of these protocols, which were mandated under Senate Bill 553 in 2023, include emotional recognition technologies, Myers said.
“Enhanced workplace security tools that have AI capabilities help to close the gap in the time it takes to issue an alert, keeping employees and the workplace safe,” Myers said in an email statement. “The same workplace security tools employers use to comply with SB 553 would be prohibited if AB 1883 becomes law, as many of these tools are capable of ‘emotional recognition.’”
CalChamber and other opponents to AB 1883 successfully lobbied to removed prohibitions on AI facial recognition technology and gait analysis in workplaces. They also struck down a clause that barred employers from using AI to obtain information about workers’ protected traits, which include race, disability, sex, age and religion.
However, proponents of workplace surveillance say that the amendments are not enough to secure their support for the bill. While CalChamber appreciates amendments made to remove prohibitions on facial and gait recognition technology, Myers said, the bill still limits their use.
“AB 1883 still has a broad prohibition on workplace security tools that are capable of making inferences about emotional recognition or neural data,” Myers said. “There are no exceptions in these provisions for workplace security tools that are deployed for safety purposes.”
Read more Florida State Challenging Dynamic Desir Twins To Reach Another Level