New Jersey employers would have to warn workers about AI layoffs

A proposal in New Jersey would require employers to warn the state before artificial intelligence, or AI, is expected to reshape a workplace on a large scale. Under the bill, any employer that adopts, expands or materially changes an AI system would have to notify the Division of Wage and Hour Compliance in the Department of Labor and Workforce Development at least 90 days before the change takes effect if the employer reasonably expects the system to eliminate, reduce or materially restructure 25 or more jobs within a 12-month period.

That is the central idea behind the New Jersey Artificial Intelligence Workforce Transition Act. Instead of treating major AI-driven staffing changes as a private internal decision, the proposal would turn them into something the state can see coming. Workers, labor officials and the public would get advance notice before the changes land, which could matter for paychecks, household budgets and the communities tied to a large employer.

When a business has to file notice

The bill does not cover every new tool or software update. It is aimed at AI changes that employers believe will have a meaningful effect on staffing. A company could fall under the rule if it adopts a new AI system, expands an existing one or makes a material modification and expects that move to change the size or structure of its workforce in the near term.

The threshold is set at 25 positions over a 12-month period. That makes the proposal broad enough to cover a significant workplace shift, but not so broad that it would sweep in minor automation changes. In plain language, the bill is trying to distinguish between routine technology upgrades and changes large enough to matter for a local labor market.

If an employer misses the deadline, the bill sets civil penalties based on how late the notice is. The farther in advance the notice is filed, the lower the penalty. Notices filed between 60 and 89 days before implementation would bring one penalty level. Notices filed between 30 and 59 days before implementation would bring a higher one. Notices filed less than 30 days before implementation would bring the highest penalty in the structure described in the bill text.

The bill also gives workers a way to complain. A worker who believes an employer broke the notice requirement could file a written complaint with the division. The division would have to acknowledge receipt within 10 business days and complete a preliminary review within 45 days. The proposal also bars employers from retaliating against a worker who files a complaint, takes part in an investigation or testifies in a proceeding. A worker who faces retaliation could seek reinstatement, back pay, compensatory damages and reasonable attorneys' fees in Superior Court.

The notices would not stay hidden

The proposal does not leave the warning inside a government file. Once a notice is filed, the Department of Labor and Workforce Development would have to post it on its website within 30 days. The posting would need to be searchable and machine-readable, which makes it easier for the public to find and use.

That detail matters because searchable notices could let people see where AI is changing work before the effects show up in jobless claims or local layoffs. Reporters, union leaders, workforce planners and community groups could track patterns by employer and by place. For workers, the notice could provide an earlier sign that a job change is coming and a little more time to plan.

The bill also gives the division another tool. It could cross-reference the employer notices with unemployment insurance claims data. That could help the state see whether reported AI changes line up with later job loss or broader shifts in employment.

Help for workers who are displaced

The bill is not only about notice. It also lays out support for workers who lose jobs because of AI. An AI-displaced worker who is otherwise eligible for State unemployment insurance benefits would receive an extended benefit period of 26 additional weeks beyond the standard maximum.

The proposal also includes a wage support piece. An AI-displaced worker who moves into a new job could receive payments equal to half of the wage difference between the old job and the new one, up to a weekly cap of 400 dollars and for as long as 52 weeks. That kind of support could soften the immediate hit when a worker has to take a lower-paying role after a layoff or restructuring.

Training is part of the package too. The bill would make an AI-displaced worker eligible for an individual training grant of up to 6,000 dollars per fiscal year for two fiscal years. The money could go toward tuition, fees and materials. In practice, that would help cover retraining costs for workers who need a new credential or a new set of skills to move into a different field.

The proposal also directs the state to prioritize workers in occupations it deems highly vulnerable to displacement. Within that group, the state would weigh factors such as liquid savings, age, how dense the local labor market is and how easily a worker's skills can transfer to another job. That is an attempt to send the most help to people who may have the least room to absorb a sudden disruption.

The bill also tries to push employers toward retention

The workforce transition plan does not only rely on penalties and benefits. It also creates tax credits meant to reward businesses that keep people working while they adopt AI. One credit would go to taxpayers that maintain full-time equivalent headcount above a baseline while implementing an AI system. The credit would be tied to the number of workers retained above that baseline.

Another credit would cover the direct cost of retraining existing employees for new roles within the same company, including positions created by or adjacent to an AI deployment. That credit would cover tuition, fees, instructor costs and approved training materials. The bill also includes a separate credit for businesses that establish or expand operations in a designated artificial intelligence economic opportunity zone and create net new full-time human jobs there.

Taken together, those provisions suggest the proposal is trying to do two things at once. It would make big AI-related staffing shifts more visible, but it would also encourage companies to keep people on payroll, retrain current staff and create new human jobs when they bring in automation. The state would be trying to steer companies away from a simple cut-and-replace approach.

The bill also reaches beyond large employers. It encourages county community colleges to offer stackable, stand-alone credentials lasting four to 18 months in high-demand fields identified by the State Employment and Training Commission. It would also direct the labor commissioner to designate AI-adjacent occupations as priority sectors under the existing Growing Apprenticeship in Nontraditional Sectors, or GAINS, grant program. In plain terms, the proposal would push training providers and apprenticeship dollars toward the kinds of jobs likely to sit next to AI rather than disappear because of it.

A five-year test of how New Jersey responds to AI

The act is written as a five-year measure. That gives the state a limited window to see whether early notice, public reporting and worker supports change how employers handle AI-driven restructuring. If the system works, New Jersey could have a clearer picture of where AI is affecting jobs, which workers are most exposed and where retraining dollars are going.

The department would also be required to publish quarterly measures of AI job displacement rates by industry sector, occupation type, county and ZIP code. On top of that, the commissioner would have to publish an adaptive capacity index by occupation and ZIP code. That index would estimate how able workers are to navigate job displacement based on factors like savings, age, local labor market density and skill transferability.

For a general audience, that means the state would not just be counting layoffs. It would be trying to map vulnerability. Officials could see where workers may have trouble bouncing back and where support might need to be more targeted.

In the end, the bill treats AI as both a business tool and a labor issue. It does not try to stop companies from using it. It tries to make the consequences visible earlier, give workers more protection if they are displaced and steer the state toward training and retention before the damage is done.