Equitable Access to AI Tools

SECTION 1. TITLE.

This Resolution may be referred to as the “Equitable AI Access Resolution.”

SECTION 2. DEFINITIONS.

SGA — The term “SGA” shall refer to the Student Government Association of The Johns Hopkins University.JHU, Hopkins - The terms “JHU” and “Hopkins” shall refer to the Johns Hopkins University.Advanced capabilities — The term “advanced capabilities” shall refer to features beyond basic chat that are provided within institutional data protections and that continue to evolve. They currently include access to current frontier models; web search and deep research; file analysis and persistent projects with memory; connections to University applications and data; agentic task delegation and scheduled automation; and coding tools.Native client — The term “native client” shall refer to an AI provider’s own applications, including its web, desktop, mobile, and command-line tools, as distinct from access to the provider’s models through an application programming interface (API).Baseline seat — The term “baseline seat” shall refer to an individual license in a native client that includes the advanced capabilities at standard usage limits.Students — The term “students” shall refer to the undergraduate students of the Krieger School of Arts and Sciences and the Whiting School of Engineering, whom the SGA represents.Governance Committee — The term “Governance Committee” shall refer to the University’s AI Teaching and Learning Governance Committee.

SECTION 3. FINDINGS AND PURPOSES.

AI tools have become part of how students learn, conduct research, and prepare for careers. Used responsibly and within instructors’ course policies, they help students work through difficult material, analyze data, write and debug code, and build skills that graduate programs and employers now expect.Much of the value of these tools now lies in the environment around the model, not the model alone. AI providers’ native clients offer advanced capabilities, and providers continue to add new ones.HopGPT, which the University provides to the entire Johns Hopkins community at no cost, gives secure access to frontier models from several providers through their APIs, is approved for sensitive data, including protected health information, and continues to add models and features. Because it reaches these models through APIs, it cannot fully provide the environment that native clients offer, such as projects and memory, connectors, agentic task delegation, deep research, scheduled tasks, and desktop and command-line tools.Through the University, students can obtain a native client with advanced capabilities only through a paid license billed to a departmental budget with a fiscal-year commitment. As a result, students who want these capabilities within the University’s data protections must belong to a lab or department that funds them; others pay for personal subscriptions, which fall outside those protections, or go without.As of October 2026, a growing number of peer institutions provide a native client to all of their students at no cost:The University of Chicago provides every student a Claude Enterprise standard seat, including Claude Code and Cowork.Syracuse University provides every student Claude Enterprise, with Claude Code and Cowork available by request after a short training.Northeastern University provides every student a standard Claude plan and grants Premium seats, which add Claude Code, by approved request.Duke University provides all university students ChatGPT Edu, including Codex.Dartmouth College provides every student Claude for chat; Claude Code is available to students only in classes whose instructors request it.Other close peers, including the University of Pennsylvania and Columbia University, charge for access or fund it through individual schools and departments. Under those models, as under the University’s current one, a student’s access depends on whether a department pays for it.A centrally funded baseline seat is the only one of these models in which access does not depend on a student’s resources or affiliation. At Hopkins, where undergraduates take part in research across disciplines, it would also bring students who now rely on personal subscriptions within the University’s data protections.The guiding principles adopted by the Governance Committee include equity.The purpose of this Resolution is to state the SGA’s formal opinion that equitable access to advanced capabilities is a student need, and to recommend a direction to the University without specifying a vendor.

SECTION 4. RESOLUTION

The SGA affirms that every student should have equal access to advanced capabilities, regardless of their ability to pay or their research affiliation.The SGA recommends that the University provide every student a centrally funded baseline seat in a leading AI provider’s native client, such as OpenAI’s or Anthropic’s, that includes the advanced capabilities, under an institutional agreement with the University’s data protections. The SGA does not endorse a specific vendor.The SGA recommends that access beyond the baseline seat, such as higher usage limits or the most resource-intensive models, be granted by request only, based on academic or research need and through a single University-wide process, rather than offered for purchase, so that upgrades do not recreate the gap this Resolution seeks to close.The SGA recommends that any new tool be optional for students, that its use remain governed by instructors’ course policies and the University’s guidelines for responsible AI use, and that HopGPT remain available as the University’s secure, multi-model platform, including for work involving protected health information.The SGA encourages the University to extend the same access to all other Johns Hopkins students, graduate and undergraduate, across its divisions.If the University provides such access, the SGA commits to help drive student adoption by promoting the tools through SGA channels, supporting student training, and communicating clearly how University-provided AI tools handle student data, including the difference between security logging and monitoring.

SECTION 5. ACTION ITEMS.

To be done by the Chair of Academic Affairs within one week of adoption: Transmit this Resolution to Vice Provost Janet Schreck and Dr. James Diamond, co-chairs of the Governance Committee, and to Marc Overcash, Vice President and Deputy Chief Information Officer.To be done by the Chair of Academic Affairs before February 28, 2027: Meet with the Governance Committee co-chairs and Johns Hopkins Information Technology to follow up on the University’s response, and report back to the Senate.The Bill is hereby ADOPTED/LOST by a vote of X to X with X abstentions.Upon passage, we have hereunto set our hands and caused the seal of the Senate to be affixed to this bill indicating the Senate’s assent and thus officially declaring the opinion of the sga.

NAMEStudent Body President of the Johns Hopkins SGA

NAMEStudent Body Vice President of the Johns Hopkins SGA

NAMEPresident of the Senate of the Johns Hopkins SGA

NAMEStudent Body Secretary of the Johns Hopkins SGA

NAMEStudent Body Treasurer of the Johns Hopkins SGA

NAMEStudent Body Chair of Programming for the Johns Hopkins SGASignatures of Ceremony:

NAME, POSITION

NAME, POSITION

NAME, POSITION

NAME, POSITIONNames of Petitioners:NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)NAMEPOSITION(S)