Pollster: ‘People are ready for Massachusetts to act first’
BOSTON, July 20, 2026…..The Senate is poised to embrace guardrails on major artificial intelligence developers this week and voters in a new poll say they support the type of framework embedded in the economic development bill and are concerned about AI-induced job losses.
While nearly 40% of Bay Staters use AI tools multiple times each week, the majority (65%) of individuals say they are more concerned about the potential dangers of AI compared to the 16% who are more excited about the prospect of AI helping humanity, according to a new MassINC Polling Group survey sponsored by AI developer Anthropic.
Poll Finds Strong Support for AI Safety Rules
“There’s not hesitation to use AI systems for people’s own lives,” Steve Koczela, the polling group’s president, told the News Service after a poll release event at the UMass Club. “It’s just they want to make sure somebody’s keeping an eye on what’s going on.”
The poll queried 1,264 registered voters and had a margin of error of plus or minus 3.7 percentage points. The survey was taken between June 23 and July 1. MassINC announced the event on June 29, and it was initially slated for July 8. On July 1, MassINC changed the event date to July 20. Senate Democrats released their AI regulation proposals on July 16.
Without regulations and legal safeguards, the poll found over 80% of respondents think it’s very likely or somewhat likely that AI systems “could create serious risks, such as enabling chemical, biological, radiological, or nuclear harms, large-scale cyberattacks, or other events causing widespread harm.” Seventy-five percent of respondents strongly or somewhat support a proposal requiring the largest AI companies to assess catastrophic risks, publish safety plans, report serious safety incidents to the attorney general, and face penalties for non-compliance.
The findings, Koczela said, mean the Senate Ways and Means Committee economic development bill (S 3178) is popular among voters. Sixty-six percent of respondents say Massachusetts should play a “leading role” when it comes to setting safety parameters for AI systems, rather than waiting for the federal government to act first, the poll found.
“This isn’t one of the ones where people want everybody to go at the same time, or don’t want Mass. to go unless Connecticut, New Hampshire go,” Koczela said. “People are ready for Massachusetts to act first.”
Senate Bill Targets Frontier AI Developers
Slated for floor consideration on Wednesday and Thursday, the Senate Ways and Means legislation requires large frontier AI developers to create safety frameworks aimed at mitigating risks of potential catastrophes and critical safety incidents. The attorney general would be allowed to sue developers for violating the safety requirements.
The bill would also create a commission to study AI safety and regulatory issues, such as tasking third-party auditors to verify whether AI developers comply with safety protocols.
The majority of voters support legislation that requires large AI companies to publicly explain how they handle major risks, update their safety plans at least once a year and explain any major changes, regularly check whether their systems could inflict “very serious harm,” check whether safety steps are working, alert the attorney general within 15 days of finding serious safety problems, report an “urgent problem” within 24 hours if people could be killed or seriously hurt, pay up to $1 million per violation, and protect AI company employees from any retaliation for reporting safety problems, according to the poll.
The Senate bill authorizes $75 million to support the development and application of AI. Half of poll respondents strongly or somewhat support the creation of an AI innovation fund, while 31% are strongly or somewhat opposed and 18% didn’t know or didn’t answer.
Experts Say More Regulation May Be Needed
John Basl, an associate professor of philosophy at Northeastern University who specializes in the ethics of emerging technology like AI, said the Senate’s focus on catastrophic safety risks is “important” but represents a “pretty narrow slice of the pie of the risk we’re facing.”
“One thing we know is that once you have a law in place, that can sort of concretize that voters could start saying, ‘OK now this technology is safe. I have to worry about it less,'” Basl said. “Which provides less motivation for the sort of more invidious risks or invasive risks. So like discrimination concerns, privacy concerns, work concerns, labor concerns, displacement concerns — all of those things need to be a part of a robust regulatory strategy.”
A conference committee is hashing out consumer data privacy legislation, with the House and Senate taking different approaches on just how narrowly to limit the collection of personal data and whether to allow for private lawsuits. Kade Crockford, director for technology and justice programs at the ACLU of Massachusetts, said data privacy and AI issues “absolutely” go hand in hand.
“We at the ACLU and our coalition, we’re focused on data privacy this session, which is really the foundation for any regulation on AI,” Crockford said. “Once we can get through this process, and we hope get a good bill signed by the governor, next session we’re going to be looking more closely at artificial intelligence and anti-discrimination.”
Crockford said existing laws prohibit discrimination in areas like hiring, employment and housing.
“The problem is that many of these AI systems are not subject to rigorous pre- and post-deployment testing that would ferret out this discrimination,” Crockford said.
AI Job Loss Emerges as Top Public Concern
The House Ways and Means Committee dropped AI safety language from its economic development bill, and the discrepancy could be a flashpoint in future negotiations. Anthropic had endorsed the safety protocols, The Boston Globe reported last month.
The MassINC poll found 32% of respondents were most concerned about AI-induced job losses and economic impact, while 16% pegged loss of skills and overreliance on AI. Others worries revolved around misinformation, environmental impacts, misuse and abuse of power, and privacy and security.
“Anthropic is looking into a study for unemployment insurance related to AI job loss and job displacement. I think that’s something we could also encourage other companies to do,” said Beauclarine Thomas, Anthropic’s state and local government relations lead.
Thomas also touted the new Claude Corps program, which she said “is geared towards individuals with two years or less of work experience, where they’re able to learn how to use AI and be trained, and to go directly and work with a nonprofit to help them with their modernization efforts.”
Basl said Northeastern students are struggling to secure entry-level co-ops with the rise of AI. Others are “very anxious” about pursuing degree paths like computer science, he said.
“We’ve had a huge uptick in the number of philosophy majors because they see that as a viable skill they can supplement this with,” Basl said.
