For centuries, the period known as the Dark Ages was defined by stagnation — a time when innovation slowed and institutions resisted change. New ideas were often viewed with suspicion, and preserving the status quo took precedence over progress.
Some proposals now moving through the Connecticut General Assembly reflect a similar mindset.
Two bills would significantly restrict how businesses adopt new technology. One targets artificial intelligence. The other regulates grocery store self-checkout machines. Taken together, they signal an increasingly aggressive effort to slow automation rather than prepare workers for it.
The first proposal, Senate Bill 435, would regulate how employers use artificial intelligence in hiring and workplace management. While framed as protecting workers from biased algorithms, the bill’s most consequential provision requires employers to bargain with unions over the use of AI technology.
Under the proposal, companies would have to notify unions before deploying artificial intelligence systems that affect unionized employees and negotiate over the purpose and impact of the technology. More importantly, employers could not use AI in ways that “modify or impair” a collective bargaining agreement, including technologies that reduce employee hours or automatetasks previously performed by workers.
In practice, introducing AI in unionized workplaces could require union approval.
The bill also requires companies to hire a state-approved auditor to test AI systems for bias before deployment. If disparate impact is found, the system cannot be used until the Labor Commissioner signs off on corrective measures. Businesses would effectively need a government permission slip to deploy their own software.
The legislation goes even further by limiting how state agencies may use artificial intelligence, requiring legislative authorization before purchasing or deploying many AI systems. Rather than encouraging innovation, the default posture treats AI as suspect unless expressly approved.
At the same time, Senate Bill 438 would regulate self-checkout machines in grocery stores.
The bill would require at least one staffed checkout lane for every two automated stations, cap self-checkout machines at eight per store, and mandate a dedicated employee to monitor every two machines — not just during busy hours but at any time the machines are operating. A store open at 2am with three customers would still be required to staff four manned checkout lanes.
Those employees would not even be allowed to perform other duties while assigned to monitor self-checkout.
And if a store falls short? The bill creates an enforcement mechanism that allows customers to file complaints with the Labor Commissioner reporting that their grocery store has too many self-checkout machines. Connecticut would, in effect, create a hotline for shoppers to report insufficient cashiers.
Viewed together, these bills reflect a broader strategy: using regulation to slow the adoption of automation in the workplace.
Organized labor has historically resisted technological changes that disrupt existing job structures. From manufacturing automation to self-service gas stations, new tools often face opposition when they alter how work is performed.
The concern for workers is understandable. But the policy response matters.
Economic growth, not stagnation, has historically provided the strongest protection for workers. In expanding economies, businesses compete for employees, wages rise, and new opportunities emerge.
Policies that slow innovation do not eliminate change. They delay it — often at the cost of investment, competitiveness, and job creation.
If Connecticut makes it harder to deploy new technology, businesses will not abandon innovation. They may simply choose to invest elsewhere.
The technology will arrive. The question is whether the jobs and growth arrive with it.
Artificial intelligence is already transforming how companies operate, from scheduling and logistics to hiring and data analysis. These tools allow companies to operate more efficiently and enable smaller firms to compete.
If deploying such technologies requires months of negotiations or new regulatory hurdles, companies won’t sit around waiting for permission. They will invest where innovation is easier.
If lawmakers want to help workers adapt to automation, workforce training, education, and skills development are forward-looking solutions. Attempts to freeze existing workplace structures through restrictive mandates risk doing the opposite.
The debate now facing Connecticut is not about whether technology will advance. It is about whether the state will compete in a modern economy, or make itself less attractive for innovation and growth.
Trying to legislate against technological change has rarely worked. The long-term cost of delay is often borne not by machines — but by workers and communities left behind.
