The Connecticut Senate passed Senate Bill 5 on Tuesday by a vote of 32 to 4. The bill requires employers to notify workers when AI is used in hiring, promotion, or termination decisions.

It prohibits employers from deploying AI tools that produce discriminatory outcomes. It funds an AI literacy program for workers and small businesses and creates a state regulatory sandbox where companies can test new AI tools under oversight.

The bill’s sponsor, Senator James Maroney, was direct about the federal vacuum it fills: “We can no longer wait for DC to act.” The bill now moves to the House before a May 6 deadline.

Also listen on Apple Podcasts

Here’s why the timing matters: The White House published a “National AI Legislative Framework” in March calling for a unified federal approach to AI regulation. The document contains no binding obligations on employers. It signals that the administration wants to preempt state laws while producing no federal replacement.

Connecticut’s Senate passed its worker protection bill two days after that framework circulated widely. The collision between a federal preemption posture and accelerating state legislation is now the central regulatory story in AI and employment.

Workers in Connecticut may know by May 6 whether their employer must tell them an algorithm screened them out. Workers in most other states have no such right.

The urgency behind Connecticut’s bill becomes clearer when set against what is happening to the managers who are supposed to be implementing AI at the team level.

Fortune reported on April 7, drawing on Bureau of Labor Statistics data, that the average American manager now oversees 12 direct reports. That figure has nearly doubled since Gallup began tracking it in 2013.

Companies have spent three years gutting middle management and leaving whoever survives with a dramatically larger portfolio of people. Gartner found that 75% of HR leaders believe managers are already overwhelmed by their expanding responsibilities.

Yesterday’s AI Labor Report established that workers are 8.7 times more likely to say AI has transformed their work when their direct manager actively supports adoption. The managers capable of driving that transformation are being eliminated or buried under workloads that make mentorship and coaching impossible.

A Harvard Business Review piece published this month, co-authored by Wharton researchers, puts a name on the resulting organizational failure. Senior executives experience AI as a strategic advantage. Middle managers confront its actual failures inside real workflows, under real constraints, without adequate support or time.

AI initiatives stall because the people charged with making them work see a fundamentally different reality than the people who approved the budget.

The Wharton authors are direct: companies are treating AI adoption as a top-down mandate and skipping the operational burden in the middle. The result is billions in investment producing the outcome Gallup The AI Labor Report discussed this week — widespread individual use, almost no organizational transformation.

Into that gap steps Jack Dorsey. In a post published April 1 with Block’s lead independent director Roelof Botha,

Future Forwarded is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Dorsey argued that AI should replace middle management entirely.

The post describes AI handling coordination, task assignment, project tracking, and real-time business insights — the core functions that management layers have historically performed.

Block cut 40% of its workforce before publishing the piece. Dorsey’s post is one of the first explicit executive statements treating middle management elimination as an organizational design goal rather than an unfortunate side effect of restructuring.

The three stories are connected.

The managers who would implement AI humanely are being removed.

The executives who remain are treating AI as a mandate from above.

The workers who bear the cost of both decisions are now waiting to see whether state legislatures will step in with the transparency and protections that the federal government has declined to provide.

So the AI deployment battleground narrows from ructions between labor-and-industry to workers-and-management.

Check out the Future Forwarded Collections of its most popular Substack series. Each collection is $9.95 flat fee, in PDF or ePUB formats. No subscription required. Each collection takes about two- to three-hours to read.

Get full access to Future Forwarded at futureforwarded.substack.com/subscribe

Get full access to The AI/Labor Report at ailabor.substack.com/subscribe