FAQ

The Lab is in a pilot phase with limited resources. Access is open to all Graduate Center students, faculty, and staff, and to selected faculty and staff from other CUNY campuses whose work aligns with the Lab's research and pedagogical framework. Students at other CUNY campuses, graduate and undergraduate, can only join if part of a class pilot or project.

Eligibility does not guarantee approval. Apply for access.

Every Lab tool uses CUNY Login. Open the tool and you will be sent to the CUNY sign-in page. In the CAIL Sandbox, choose Continue with CUNY Login and leave the email and password fields blank.

Signing in confirms who you are; it does not grant access on its own. Your application or class invitation must be approved first. See How to Sign In to the CUNY AI Lab.

This is caused by saved data for CUNY's sign-in site (ssologin.cuny.edu) in your browser.

1. Close the stalled tab.
2. Open the tool, or your class invitation link, in a private or incognito window, or in a different browser, and sign in there.
3. If that works, clear cookies and site data for ssologin.cuny.edu only in your regular browser, then sign in again. You do not need to clear anything else.

How to clear data for just that site:

• Chrome: open chrome://settings/content/all (Settings → Privacy and security → Third-party cookies → See all site data and permissions), search for cuny.edu, and choose the trash icon next to ssologin.cuny.edu. Chrome help.
• Firefox: Settings → Privacy & Security → Cookies and Site Data → Manage Data, search for cuny.edu, select ssologin.cuny.edu, and choose Remove Selected. Firefox help.
• Safari on Mac: Safari → Settings → Privacy → Manage Website Data, search for cuny.edu, select the entry, and choose Remove. Safari help. On iPhone or iPad: Settings → Safari → Advanced → Website Data. iOS help.

If CUNY Login completes but the Lab says your access is pending or not active, clearing browser data will not help. Your application or class invitation still needs approval.

One application covers the CAIL Sandbox, personal API keys for the Lab's model endpoint, Agent Studio, Site Studio, and Media Tools. You do not need to apply separately for each tool.

An active account lasts one semester by default. That can be changed: if your project runs longer, email ailab@gc.cuny.edu to ask for an extension. Class access follows the approved course dates.

Apply for a class on the Request Access page. You do not need existing Lab access to apply. Once the class is approved, you receive an invitation link to share with your students. Each student opens the link, signs in with CUNY Login, and their class access is active immediately. Students in a class pilot do not submit individual applications.

Class access lasts for the approved course dates.

The standard allowance is $5 per month of model use. Your Dashboard shows the allowance and reporting period currently assigned to your account.

We welcome opportunities to raise that allowance. If you or your program have funding you would like to allocate to Lab use, please contact us to explore a collaboration.

Sign in and open your Dashboard. It shows your current reporting period, how much you have used, your limit, and an estimate of what remains. The estimate comes from provider analytics and can lag a short while behind your most recent requests.

Model requests pause when you reach your limit. Your Dashboard shows your current reporting period. If you need more for a specific project or course, email ailab@gc.cuny.edu to ask about an exception.

The Model Registry lists every model currently available through the Lab, with its capabilities and the exact API ID to use. The list is read live from the Lab's gateway, so it is always current.

The Lab runs on open-weight models: models whose weights are published, so they can be inspected, hosted by different providers, and audited. They cost far less to run than proprietary models and now perform close to them on most academic tasks. Read Running on Open Weights.

A free page at tools.ailab.gc.cuny.edu/ai-disclosure that helps you write a statement describing how AI was used in a piece of work, following the GAIDeT taxonomy. The statement is generated in your browser; nothing you type is uploaded or stored.

Where it lives. The CAIL Sandbox runs on Amazon Web Services in the US East (N. Virginia) region. The Tools Portal apps (Media Tools, Agent Studio, Site Studio, and sign-in) run on Cloudflare Workers, with files stored in Cloudflare R2.

What is stored.

• Sandbox chats and uploads are saved to your account until you delete them. There is no automatic expiry. A Temporary Chat is never saved.
• Media Tools deletes uploaded audio, video, and images within one day. Finished transcripts stay in your account until you delete them, with word-level timings stripped before storage.
• Agent Studio and Site Studio keep your workspaces, projects, and files until you delete them.
• API keys are shown once and stored only as a hash.
• Sign-in data. Your CUNY Login identity stays in the Lab's sign-in service. Individual tools receive an opaque ID rather than your CUNY username.
• Logs record events, not content. The Lab's model gateway stores no prompts or responses.

What model providers see. Prompts and files are sent to a model provider only for the time it takes to generate a response. The Lab uses two providers, both configured for zero data retention, and only enables models that permit it.

• Cloudflare Workers AI. Cloudflare does not use customer content to train AI models or to improve its services, and does not make it available to other customers. See the Workers AI data usage page, the Developer Platform terms, and the Enterprise Subscription Agreement.
• Amazon Bedrock. By default, Bedrock does not store model inputs or outputs, no AWS operator can access them, and they are not shared with model providers or used to train models. Content stays in the AWS region where it is processed. See Data protection, Data retention, Abuse detection, the Bedrock FAQs, and the AWS Data Privacy FAQ.

What you should not upload. Confidential, FERPA-protected, HIPAA-regulated, or personally identifying information.

The Lab's Terms of Use and Privacy Policy are under review and will be published here soon.

No. Do not enter confidential, FERPA-protected, HIPAA-regulated, proprietary, or personally identifying information into any Lab tool.

If your research involves audio, images, interviews, archives, or institutional data that should not leave the Graduate Center, contact us about on-premises processing. We handle those requests by consultation.

They are under review and will be published here soon.

Email ailab@gc.cuny.edu, or use the consultation and bug-report forms on the Contact page. The Sandbox documentation answers most questions about the chat platform, and the Events page lists upcoming workshops.

Given our limited capacity, the Lab cannot provide direct support to classes beyond very specific cases. When you sign up for class access, you also sign up to provide technical support for your students.

We expect instructors to be the first layer of troubleshooting, for example the blank sign-in screen described above, and to contact the Lab only after those options have been exhausted.

Instructors are also responsible for helping students manage their allowance so that no one reaches their limit before completing classroom assignments. The standard allowance is $5 per month, but each student should check the allowance and reporting period shown on their Dashboard. Plan assignments accordingly.

A yearly fellowship for Graduate Center master's and Ph.D. students to work with the Lab. Calls are announced on the blog; see the 2026–27 call.