Promoting economic development with emerging artificial intelligence models and safety
The introduction of S37 could significantly reshape existing state laws regarding technology development, adding a layer of regulation and oversight specific to AI models and their deployment. The bill amends Chapter 29 of the General Laws and introduces a new chapter, 93M, addressing the operational requirements for developers of AI models. These new regulations ensure that developers must adhere to safety and security protocols, conducting regular assessments to mitigate risks associated with AI. For example, the bill entrusts the attorney general with the oversight of compliance and safety incident reporting, to protect public safety while fostering innovation.
Senate Bill S37 seeks to promote economic development in Massachusetts by facilitating growth in artificial intelligence (AI) sectors. It establishes the Massachusetts Artificial Intelligence Innovation Trust Fund, managed by the Secretary of Economic Development, designed to provide grants and financial support to companies involved in the development and deployment of AI technologies. The bill aims to encourage entrepreneurship and support research initiatives related to AI, while requiring the engagement of private sources for matching funds before financial disbursements from the fund are made. The fund will also channel investments towards education and partnerships with research institutions to build a robust AI ecosystem in the state.
While there is a clear push towards innovation, the bill has raised concerns regarding the balance between economic interests and public safety. Key points of contention include the definitions of 'critical harm' and 'covered models,' which may raise compliance difficulties for developers. Moreover, the requirement for regular third-party compliance investigations could be viewed as cumbersome by industry stakeholders who fear that excessive regulation may stifle innovation. Furthermore, the expansion of responsibilities for AI developers pertaining to the safeguarding of their models against misuse is significant, raising questions about the feasibility and implications for smaller AI entities who may lack the necessary resources to comply fully.