Executive Office of Technology Services and Security (EOTSS)

EOTSS deployed a ChatGPT-powered AI assistant for Massachusetts state employees, making it available to nearly 40,000 workers across the executive branch through a phased rollout beginning in February 2026. The state says this makes Massachusetts the first state to adopt the tool enterprise-wide.

Following a competitive procurement process, EOTSS contracted with OpenAI (via vendor Carahsoft Technology Corporation) to provide a ChatGPT-based assistant to state employees. The rollout began with EOTSS staff and is expanding to other secretariats and agencies over subsequent months. The tool operates within a secure, walled-off environment where employee inputs are not used to train public AI models. Employees are directed to use it for tasks such as drafting documents, summarizing reports, and researching work-related topics. Use is voluntary and governed by terms and conditions set by the EOTSS Privacy Office. The state is also offering optional training programs to help employees use the tool responsibly.

EOTSS uses a formal Generative AI Policy that sets mandatory requirements for how all Massachusetts executive branch agencies may develop and use AI tools, including requirements for AI inventories, risk evaluations, and compliance with ethical standards. The policy was first published and has been updated since 2024.

The Enterprise Use and Development of Generative AI Policy, issued by EOTSS, establishes minimum requirements for all Commonwealth agencies and offices using or building generative AI. Key requirements include: completing an AI Project Risk Evaluation before procuring or deploying any AI system (submitted to the AI Center of Excellence for review); creating an agency-level Generative AI Inventory; and operating AI within the EOTSS AI sandbox or a comparably secure environment. The policy also mandates that human employees remain accountable for AI-supported decisions and must review and validate AI outputs. EOTSS updates the policy periodically in response to evolving technologies and guidance from the AI Strategic Task Force.

EOTSS offers state agency employees a secure AI sandbox — a walled-off testing environment — that lets them experiment with generative AI tools without risking exposure of sensitive government data to public AI systems. The sandbox was established in 2024 and serves as the foundation for EOTSS's AI Center of Excellence.

The AI sandbox is a cordoned-off, cloud-based environment where state employees and agency teams can test generative AI use cases without their inputs being used to train public AI models. The sandbox was initially developed with assistance from Northeastern University's Burnes Center (through the GENIE project) and is administered by the EOTSS AI Center of Excellence, established in 2024. All executive branch project teams intending to use or procure AI are required to complete an AI Project Risk Evaluation reviewed by the Center. The sandbox preceded the broader ChatGPT enterprise rollout and served as the testing ground for many of the agency-specific AI use cases now moving into production.