Axios Live Brings Boston AI Forum to Massachusetts as State Weighs Safety Rules

Axios Live is bringing its AI Across America: What's Next for Boston forum to Boston on Sept. 29, putting Massachusetts' AI ambitions, cybersecurity exposure and safety rules into one public program. The event arrives as the state is trying to turn its research and startup base into applied AI infrastructure while voters and lawmakers press for stronger controls on frontier systems. The program says streamed conversations begin at 8:30 a.m. ET, while the Axios Live listings place the in-person event at Hotel Commonwealth at 8:00 a.m. EDT. Axios business editor Dan Primack, senior business reporter Hope King and Boston reporter Steph Solis are scheduled to speak with Bain Capital Ventures partner Ajay Agarwal, Massachusetts state Sen. Barry Finegold, Recorded Future CEO Colin Mahony and Anthropic Head of North America Government Affairs Brian Peters about AI adoption, emerging cyber threats, digital defenses and deployment safeguards. The local backdrop is unusually dense. The Massachusetts AI Hub, launched in 2024, describes its mission as connecting, accelerating and scaling AI initiatives for innovation, economic growth, workforce development and public good. Boston University said the Mass Leads Act provided $100 million directly to the Hub and that the Massachusetts Green High Performance Computing Center in Holyoke would support expanded AI computing access. In February, Massachusetts announced a phased, secure ChatGPT-powered assistant for nearly 40,000 executive-branch employees, and the state FAQ says the tool can draft, summarize and translate while public versions of ChatGPT, Gemini and Claude are prohibited on state-issued devices. Lawmakers are moving on the other side of the ledger. In July, the state Senate passed a $575.4 million economic development bill with $75 million to support AI development, $100 million for scientific research at public higher education institutions and $25 million for robotics, plus requirements for large AI developers to address critical safety incidents. The draft text for S.2630 defines a frontier model as one trained with more than 10^26 operations, a large frontier developer as one with more than $500 million in annual gross revenue, and catastrophic risk as more than 50 deaths or more than $1 billion in damage from a single incident. Gov. Maura Healey separately called on Sept. 18 for independent safety evaluations and stronger incident reporting in the Mass Wins negotiations. The reception reflects both demand and unease. A MassINC Polling Group survey of 1,264 registered voters, sponsored by Anthropic, found 39% use tools such as Claude, ChatGPT, Gemini or Microsoft Copilot every day or a few times a week, while 65% are more concerned about AI's dangers than excited about its benefits and 78% worry more about failing to regulate AI than about slowing innovation with rules. The same poll found 75% support, once informed, for requiring the largest AI companies to assess catastrophic risks, publish safety plans and report serious incidents to the attorney general. The cybersecurity discussion is likely to be less about science fiction than operations. Recorded Future's H1 2026 malware and vulnerability analysis said AI-enabled cyber threats became more visible, but most observed activity still augmented existing tactics rather than replacing them with fully autonomous operations. Mastercard's 2024 deal to acquire Recorded Future for $2.65 billion underscored the commercial value of threat intelligence as fraud, identity abuse and cyber defense converge. On the investment side, Bain Capital Ventures says its venture business has $10 billion-plus in assets under management and invests from $1 million of seed capital through $100 million of growth equity, while TechCrunch reported this month that BCV raised a $1.6 billion fund that will back AI-focused startups. The Boston forum therefore lands at a policy hinge point: Massachusetts is trying to claim a larger share of applied AI while making privacy, security, independent evaluation and public accountability part of the deployment stack. The unanswered question is whether the state can keep startups, public agencies and frontier developers moving quickly enough to compete while persuading skeptical residents that safeguards will work before failures become expensive or irreversible.