Week 2, March 2025: NY WARN Act AI Checkbox — Zero Companies Check It

The Checkbox Nobody Checks

In late 2024, New York State became the first jurisdiction in the United States to add an AI disclosure requirement to its Worker Adjustment and Retraining Notification (WARN) Act filings. Starting November 2024, any company filing a WARN notice — the legally required 90-day warning before mass layoffs — must indicate whether artificial intelligence was a contributing factor in the workforce reduction Source: NY Department of Labor.

The checkbox is simple. Yes or no. Was AI a factor?

After reviewing 160 WARN Act filings submitted to the New York Department of Labor between November 2024 and March 2025, AI Cuts can report:

Zero companies checked the box.

Not one. In 160 filings covering approximately 18,000 affected workers, no company disclosed that AI was a contributing factor in their layoffs.

The Filers Who Didn't Check

Among the companies that filed WARN notices in New York during this period without checking the AI disclosure box:

• Amazon — Filed notices for warehouse and logistics facility adjustments affecting 2,100 workers across New York State. Amazon's own investor communications tout AI-driven automation as a key efficiency driver in fulfillment operations Source: Amazon 10-K Filing
• Goldman Sachs — Filed notices for back-office restructuring affecting approximately 350 roles in Manhattan. Goldman has publicly stated that AI is transforming its operations, with CEO David Solomon telling shareholders that AI could replace "a significant portion of entry-level tasks" Source: Goldman Sachs Annual Report
• Meta — Filed for 280 affected roles in New York offices. Meta has been one of the most vocal companies about deploying AI across its business, with CEO Mark Zuckerberg stating that AI-generated content recommendations now drive "the majority of engagement growth" Source: Meta Earnings Call
• Citigroup — Filed for 520 roles as part of its ongoing 20,000-role global restructuring. Citi CFO Mark Mason has specifically cited AI automation of middle-office functions as a driver Source: Financial Times
• IBM — Filed for 180 roles in Armonk and other New York locations. IBM CEO Arvind Krishna has publicly committed to not replacing approximately 7,800 back-office roles that AI now handles Source: Bloomberg

The Legal Calculus

Why won't anyone check the box? The answer lies in the intersection of legal risk, PR strategy, and corporate denial.

• Wrongful termination lawsuits where employees claim they were replaced by automation
• Age discrimination claims where older workers argue they were disproportionately affected because they lacked AI skills
• Union grievances where labor contracts include provisions about technological displacement
• Future regulatory proceedings if legislatures create AI displacement compensation requirements

Employment lawyers are advising clients uniformly: do not check the box.

"No competent attorney would advise a client to voluntarily check that box," said Thomas Lenz, a partner at a major New York employment law firm, speaking to AI Cuts on background. "It creates liability with no offsetting benefit. The statute has no penalty for not checking it. It's entirely voluntary in practice, even though it's technically mandatory."

The Enforcement Gap

The New York WARN Act AI disclosure requirement has no dedicated enforcement mechanism. The Department of Labor can investigate incomplete or inaccurate filings, but:

• There is no specific penalty for failing to check the AI box
• The Department of Labor has no dedicated staff to audit the accuracy of AI disclosures
• The burden of proof would fall on the state to demonstrate that AI was a factor — which is nearly impossible without internal company documents
• Companies can easily argue that layoffs were driven by "market conditions," "restructuring," or "strategic realignment" even when AI is the underlying cause

The PR vs. Legal Disclosure Gap

This is where it gets truly absurd. Consider the disconnect:

• "AI is transforming how we operate and enabling us to do more with fewer resources" — Generic version of what virtually every Fortune 500 CEO has said in the past 18 months
• Workday CEO Carl Eschenbach: "We need to invest in AI... this means making difficult decisions about where we allocate resources"
• Klarna CEO Sebastian Siemiatkowski: "AI is doing the equivalent of the work of 700 full-time agents"
• IBM CEO Arvind Krishna: "We could easily see 30% of back-office roles replaced by AI over five years"

• "Restructuring" ✓
• "Market conditions" ✓
• "Strategic realignment" ✓
• AI as a factor? ☐ (unchecked)

The double standard is breathtaking. Companies are happy to tout AI displacement when it boosts their stock price on an earnings call. But when it comes to a legal filing that might create liability, suddenly AI had nothing to do with it.

The Sam Altman "AI Washing" Quote

OpenAI CEO Sam Altman inadvertently highlighted this hypocrisy in a January 2025 interview when he noted that companies are engaging in what he called "AI washing" — exaggerating their AI capabilities to impress investors while the technology's actual impact on their operations is more modest than claimed Source: Bloomberg TV Interview.

But the WARN Act data suggests the opposite problem: companies are understating AI's role in layoffs while overstating it to Wall Street. They want credit for AI disruption without accountability for AI displacement.

Which version is true? Is AI really transforming their workforce (as they tell investors), or is it irrelevant to their layoff decisions (as they tell the Department of Labor)? Both can't be accurate.

Other States Watching

New York's experiment — failed as it may be in practice — is being watched by other jurisdictions considering similar requirements:

• California has proposed AB-1234, which would require AI impact assessments before mass layoffs, with penalties for non-compliance up to $10,000 per affected employee Source: California Legislature
• Washington State is considering amendments to its WARN Act equivalent that would include AI disclosure with actual enforcement teeth
• New Jersey introduced a bill in February 2025 requiring companies receiving state tax incentives to report any AI-driven job displacement
• The European Union is developing workforce reporting requirements under its AI Act that would require companies deploying "high-risk" AI systems to report workforce impacts Source: European Commission
• Canada announced in January 2025 that it would study mandatory AI workforce impact reporting as part of its updated AI and Data Act Source: Innovation, Science and Economic Development Canada

What Would Work?

If disclosure requirements are going to have any teeth, they need several elements the New York version lacks:

1. Penalties for non-disclosure: Without consequences, the checkbox is meaningless 2. Independent verification: An audit mechanism that compares WARN filings against investor disclosures and public statements 3. Broader definitions: The checkbox should ask about AI as a factor in the decision to restructure, not just AI as a direct replacement — which gives companies an easy out 4. Whistleblower protections: Employees who know AI was a factor need legal protection to report inaccurate filings 5. Standardized definitions: What counts as "AI" needs to be clearly defined. Companies can argue that algorithmic automation, robotic process automation, or machine learning don't qualify as "artificial intelligence"

The Broader Transparency Problem

The WARN Act checkbox failure is a microcosm of a much larger transparency problem. We simply don't have reliable data on AI's impact on employment because:

Companies won't self-report As we've seen, companies have strong incentives to deny AI's role in layoffs and strong incentives to exaggerate AI's role in efficiency gains. The same event — cutting 1,000 jobs — is framed as "AI-powered transformation" in the investor presentation and "strategic restructuring" in the WARN filing.

Government statistics lag reality The Bureau of Labor Statistics doesn't track AI-related displacement as a category. Monthly jobs reports can't distinguish between a layoff caused by AI automation and one caused by a business downturn. The closest proxy — the BLS "mass layoff" data with reason codes — was actually **discontinued** in 2013 due to budget cuts and has not been revived [Source: BLS](https://www.bls.gov/mls/).

Researchers lack access Academic researchers studying AI displacement rely on the same public data everyone else has — company announcements, news reports, and job posting trends. They rarely have access to internal company data showing exactly which roles were eliminated because AI could do the work.

Workers are silenced Severance agreements routinely include non-disparagement and non-disclosure clauses that prevent laid-off workers from publicly stating that they were replaced by AI. Workers who want to tell their story risk losing their severance payments.

What We're Tracking

Despite the disclosure gap, AI Cuts continues to track AI-linked job cuts through alternative methods:

• SEC filings: 10-K and 10-Q reports where companies discuss AI's impact on headcount
• Earnings call transcripts: Where CEOs and CFOs discuss AI's role in "efficiency gains" and "headcount optimization"
• Job posting data: Declining postings in specific roles that correlate with AI deployment timelines
• Union reports: Labor unions, which have less incentive to minimize AI's role, provide valuable counter-narratives
• Glassdoor and Blind: Anonymous employee reports about AI replacing their teams
• WARN filings cross-referenced with AI investment announcements: When a company files a WARN notice and simultaneously announces a major AI initiative, the correlation speaks for itself

This Week's Other Developments

While the WARN Act story dominated, several other AI workforce developments occurred this week:

• Cognizant announced 3,500 layoffs globally, with internal sources telling AI Cuts that AI automation of testing and quality assurance is a primary factor Source: Economic Times
• DocuSign cut 400 roles, with CEO Allan Thygesen stating the company is "leveraging AI to automate agreement preparation and review" Source: TechCrunch
• The UK's Office for National Statistics released data showing that 34% of businesses in finance and insurance are now using AI, up from 21% a year ago Source: ONS

The Accountability Deficit

The WARN Act checkbox was supposed to be a first step toward accountability. A small, modest measure that simply asked companies to be honest about whether AI was a factor in their layoff decisions.

They couldn't even manage that.

The gap between what companies say about AI (it's transforming everything! it's the future! we're leading the AI revolution!) and what they'll legally admit about AI (it had nothing to do with our layoffs, we swear) tells you everything you need to know about the state of AI workforce accountability in 2025.

Until disclosure requirements have actual teeth — penalties, audits, enforcement — the checkbox will remain unchecked. And the true scale of AI-driven job displacement will remain hidden behind corporate euphemisms and legal maneuvering.

For workers navigating this opaque landscape, understanding your rights under state and federal WARN Act provisions is a critical first step.

---

AI Cuts obtained WARN Act filing data through public records requests to the New York Department of Labor. All filings referenced are public documents available through the NYDOL WARN Act database. Company characterizations are based on public investor communications, earnings call transcripts, and SEC filings.