Workers' Compensation: Not Fired by a Robot

NJ AFL-CIO Puts Its Weight Behind S4075, the FAIR Act, and Injured Workers Should Pay Attention

On September 24, 2026, Charles Wowkanech, President of the New Jersey State AFL-CIO, stepped to the microphone at a press conference outside the Statehouse Annex in Trenton and launched organized labor's campaign for the Fair AI Regulations at Work Act, S4075, and its Assembly companion, A4981. Speaking on behalf of roughly one million union members, Wowkanech's message was blunt. Artificial intelligence is not coming to New Jersey workplaces. It is already there, and right now the rules are being written by a small group of technology billionaires whose focus is the bottom line, not the people who will live with the consequences.

He was careful to say that labor is not against AI. He acknowledged that the technology can make work safer, more efficient, and more productive. The fight, as he framed it, is about who carries the risk when an algorithm gets it wrong. Today, he argued, the developers collect the rewards, and working people absorb the risk. His summary of the bill is the line that will likely define the campaign: “No one should be fired by a robot.”

Who Was at the Podium

Wowkanech was joined by NJ AFL-CIO Secretary-Treasurer Laurel Brennan, prime Senate sponsor Senator Andrew Zwicker (D-16), Senator Shirley Turner (D-15), and Assemblywoman Heather Simmons (D-3), who carries the Assembly bill. S4075 was introduced on May 4, 2026, and referred to the Senate Labor Committee. A4981 was introduced on May 7, 2026, and referred to the Assembly Science, Innovation and Technology Committee. Neither bill has yet been reported out of committee.

Wowkanech tied the legislative push to a second initiative. Beginning in October, the New Jersey State AFL-CIO, Rowan College of South Jersey, and the Department of Labor and Workforce Development will launch what the partners describe as the nation's first statewide AI training program built with workers and unions at the table, funded by a Department grant. Training, he said, is the first step. Passing the FAIR Act is the next one.

The Seven Ground Rules

Wowkanech walked through the bill in seven points. Each maps onto specific text in S4075 as introduced.

·       No off-the-clock spying. Section 2(d) bars electronic monitoring of an employee who is off duty, on leave, or on a meal or rest break. Section 2(e) bars requiring monitoring software on a worker's personal phone, car, or computer, and Section 2(g) bars surveillance of a worker's home or personal vehicle.
·       No AI discrimination. Section 2(j) prohibits using a monitoring tool or automated decision system to infer or rely on protected characteristics, union membership, or any trait not directly related to job performance.
·       Disclosed, safe quotas. Section 2(h) prohibits productivity quotas likely to contribute significantly to harming worker health and safety. Section 6(d) bars adverse action based on any quota that was not disclosed to the worker in advance.
·       Transparency and audits. Section 3 requires an independent impact assessment before an automated employment decision system or monitoring tool is deployed, with the report filed in a public registry kept by the Department of Labor. Section 6 requires 60 days' written notice to workers and their union, and Section 8 gives workers the right to see the data and appeal an adverse decision to a human reviewer.
·       Human oversight. Section 9 forbids any employment decision based “exclusively or determinatively” on AI outputs and requires trained human reviewers with authority to override the system.
·       Data privacy. Section 5 restricts sale or sharing of worker data, requires breach notice within 48 hours, and makes employers and vendors jointly and severally liable for resulting losses.
·       Public sector jobs. Sections 11 through 13 make public employer AI deployments that are likely to reduce employment a mandatory subject of bargaining, and guarantee displaced public employees at least one year of severance and health coverage, or two years for those with five or more years of service.

The enforcement provisions have real teeth. Section 18 makes knowing and willful violations a disorderly persons offense, with fines of $1,000 to $5,000 per affected employee for a first violation. It authorizes the Commissioner of Labor to impose administrative penalties. Section 19 creates a private right of action with reinstatement, back pay, liquidated damages of up to 200 percent, attorney's fees, civil fines of up to $10,000 per person for a first violation, punitive damages, and representative actions on behalf of similarly situated workers. Section 17 voids any waiver of the Act's rights, and Section 20 makes vendors jointly liable with the employers that deploy their tools. The bill would take effect 18 months after enactment.

Why This Matters in Workers' Compensation

The FAIR Act is drafted as an employment and labor bill. It does not amend the Workers' Compensation Act. But several provisions would directly affect how compensation claims arise, how they are investigated, and how injured workers are treated after they file.

Algorithmic quotas are an injury mechanism. In warehousing, logistics, call centers, and health care, software increasingly sets the pace. Repetitive strain and overexertion claims follow the pace. Section 2(h) would make an unsafe algorithmic quota unlawful in its own right, and the disclosure notices required by Section 6 and the impact assessments filed under Section 3 would create a documentary record of exactly what the system demanded of the worker. In a contested occupational claim, that record may become important causation evidence.
Off-duty surveillance of injured workers. Section 2(d) bars an employer, or a third party acting for it, from conducting electronic monitoring “or other surveillance” of an employee who is off duty or on leave. Read literally, a worker out on temporary disability is on leave. Whether that language would reach employer-commissioned surveillance of a claimant in a disputed compensation case is a question the bill, as introduced, does not answer. Insurance carriers are not named, but self-insured employers and their investigators arguably are. Expect this provision to draw close attention, and likely amendments, as the bill moves.
Health data and the claim file. Section 2(k) bars employers from transferring biometric, health, or wellness data to any third party unless State or federal law requires it, and from using that data in employment decisions. Section 5 separately bars sharing data collected by monitoring tools, which by definition includes video and geolocation, without the worker's uncoerced written consent, except to a court or law enforcement when required by law. Employers routinely hand accident video, telematics, and wearable data to their compensation carriers. Under the bill's text, that routine handoff may require either the worker's consent or legal process. Practitioners on both sides should watch how the final language treats claim administration.
A stronger retaliation remedy. Under N.J.S.A. 34:15-39.1, a worker fired for filing a compensation claim must prove both the claim and the retaliatory motive, as Model Jury Charge (Civil) 2.41 sets out. The FAIR Act goes further for AI-related complaints. Sections 18(c) and 19(c) create a rebuttable presumption of retaliation for any adverse action taken within 90 days after a worker complains of a violation. A worker disciplined by an algorithm after reporting an injury may have overlapping claims, and the FAIR Act presumption would be the more favorable of the two.
The right to refuse a harmful output. Section 10 protects licensed and professionally credentialed employees, nurses among them, who refuse to follow an AI output they reasonably believe will cause harm. That is a safety provision as much as a labor provision, and it recognizes that human judgment, not software, remains the last line of defense against workplace injury.
Public benefits decided by machines. The bill also covers automated systems that decide eligibility for public benefits, requiring a Department impact assessment, meaningful human oversight, and at least 10 days' advance written notice before benefits are reduced. For an injured worker whose compensation claim is denied and who falls back on State plan temporary disability or unemployment benefits, those protections are not abstract.

What to Watch

·       Committee action on S4075 in the Senate Labor Committee and on A4981 in the Assembly Science, Innovation and Technology Committee.
·       Amendments to the off-duty surveillance and data-sharing provisions, which are the sections most likely to affect claim investigation and defense.
·       The launch of the statewide AI training program in October 2026.
·       The 18-month runway. Employers already using monitoring or automated decision tools would need independent impact assessments within six months after the effective date.

Labor has now made the FAIR Act a priority, and Wowkanech made clear this is a campaign, not a single press event. For those of us who represent injured workers, the question is not whether AI will be used in the workplace. It already is. The question is whether the data it generates, and the decisions it drives, will be subject to the same fairness and accountability that the compensation system has demanded of human employers for more than a century.

Recommended Citation: Gelman, Jon L., Not Fired by a Robot, WORKERS’ COMPENSATION, workers-compensation.blogspot.com (Sept. 26, 2026), https://workers-compensation.blogspot.com/2026/09/not-fired-by-robot.html.

Sources

1. Senate Bill No. 4075, the Fair AI Regulations at Work (FAIR) Act, 222nd Legislature (N.J. introduced May 4, 2026) (Zwicker, Turner), referred to the Senate Labor Committee.

2. Assembly Bill No. 4981, 222nd Legislature (N.J. introduced May 7, 2026) (Simmons, Miller, Murphy), referred to Assembly Science, Innovation and Technology Committee.

3. Charles Wowkanech, President, New Jersey State AFL-CIO, remarks at FAIR Act press conference, Statehouse Annex, Trenton (Sept. 24, 2026) (video, Insider NJ).

4. Max Pizarro, Trenton Tackles AI in the Workplace as AFL-CIO Launches FAIR Act Campaign, Insider NJ (Sept. 24, 2026).

5. Workers Take AI Training Into Their Own Hands in New Jersey, Insider NJ (press release, Sept. 14, 2026).

6. Model Jury Charge (Civil) 2.41, Workers' Compensation Retaliation (N.J.S.A. 34:15-39.1), New Jersey Courts.

About the Author

Jon L. Gelman of Wayne, NJ, is the author of NJ Workers' Compensation Law (West-Thomson-Reuters) and co-author of the national treatise Modern Workers' Compensation Law (West-Thomson-Reuters).

Blog: Workers' Compensation

LinkedIn: JonGelman

LinkedIn Group: Injured Workers Law & Advocacy Group

Author: "Workers' Compensation Law" West-Thomson-Reuters

Blue Sky: jongelman@bsky.social

Substack: https://jongelman.substack.com/