skip to main content
Caltech
Information Management Systems and Services
IMSS  /  Services  /  Generative AI  /  AI Tools at Caltech  /  Claude AI at Caltech

Claude AI at Caltech

Claude can help with writing, summarization and analysis, coding and research support, brainstorming and productivity, and everyday work tasks.

For Caltech work, use Caltech Enterprise Claude whenever available, follow course, research, and institutional policies, and disclose AI usage when required.

Subscription Options

  • Anthropic Enterprise Standard (2x): ~2× token allocation for $240/year
  • Anthropic Enterprise Premium (5x): ~5× token allocation for $600/year
  • Subscriptions are annual only
  • You can upgrade by submitting a new request

How to Purchase

  • Submit a request at: https://access.caltech.edu/ledger_portal/ (this is also the "Tech Partner Portal" link within access.caltech, under Self Service)
  • For group purchases: include users in the request notes
  • After submitting, the Help Desk will complete account setup after billing

Getting Started: Access & Accounts

  • Log in using SSO with your Caltech email
  • You can monitor token usage in account settings

Migrating a Personal Claude Account

The directions below for exporting and importing content between Claude accounts are provided as a guidance, not a definite procedure. Outcomes will vary.

Option A — Merge (Recommended)

This option only works if your personal account is using the same email address as your Caltech account.

  1. Log in to your Caltech Claude account using the SSO option
  2. Go to Settings, Account, and choose the option to move your personal account into the organization

NOTE: The move is permanent: the personal account is closed, any paid personal subscription is cancelled with a prorated refund

Option B — Manual Export

Exporting conversation data
  1. In your personal Claude account, click on your initials in the bottom-left corner
  2. Go to Settings
  3. Click on the Privacy tab
  4. Click Export data, choose your options, and confirm

Anthropic will email you a download link once the export is processed. The export gives you your conversation data and account user data as a zip of JSON files, this export does not include Claude memory.

Importing conversation data
  1. Log in to your Caltech Claude account
  2. Paste content from the export into new chats
Exporting memory content
  1. In a chat on your personal account, paste this exact prompt below (this is Anthropic's own recommended wording, designed to extract everything cleanly):
I'm moving to another service and need to export my data. List every memory you have stored about me, as well as any context you've learned about me from past conversations. Output everything in a single code block so I can easily copy it. Format each entry as: [date saved, if available] - memory content. Make sure to cover all of the following — preserve my words verbatim where possible: Instructions I've given you about how to respond (tone, format, style, 'always do X', 'never do Y'). Personal details: name, location, job, family, interests. Projects, goals, and recurring topics. Tools, languages, and frameworks I use. Preferences and corrections I've made to your behavior. Any other stored context not covered above. Do not summarize, group, or omit any entries. After the code block, confirm whether that is the complete set or if any remain.

2. Copy the resulting code block somewhere safe. This is a good moment to review it and strip out anything sensitive you don't want carried over — especially since you're moving between a personal account and an institutional (Caltech) one.

Importing memory content
  1. In your Caltech Claude account, go to Settings, then Memory, and click Start import
  2. If you do not see Start import, look under Settings, then Capabilities, or look on the home screen for "Import memory to Claude"
  3. Paste the copied text into the import box and click "Add to memory"

This only moves memory (durable facts/preferences), not full chat transcripts or project files — those still require the separate account-level export/manual copy for actual conversation history. Anthropic explicitly flags this feature as experimental, so it may not capture everything perfectly on the first pass.

Security Best Practices

  • Use the least amount of data necessary
  • Validate all outputs (AI can be incorrect or biased)
  • Do not share outputs without review
  • Keep sensitive and non-sensitive workflows separate

Data Protection Requirements

Do NOT enter the following into AI tools, as doing so may be treated as public disclosure:

  • Student data (FERPA)
  • Health or medical data (HIPAA/PHI)
  • Unpublished research or intellectual property
  • Export-controlled data (ITAR/EAR)
  • Sensitive HR, financial, or business data

However, if you need to handle sensitive or restricted data:

  • Use browser-based Claude only (not desktop)
  • Disable Web Search:
    • Start new chat → click "+" → uncheck Web Search
  • Keep sessions isolated
  • Avoid mixing with external or unverified data

And use Cowork (Claude Desktop App) on Caltech-managed devices for:

  • Drafting emails, reports, or presentations
  • Working with non-sensitive documents
  • Performing general productivity tasks

Working with Documents

  • Use OneDrive or SharePoint via Microsoft 365 connector
  • Store documents in approved locations
  • Follow proper controls when working with institutional data

Responsible Use

  • Disclosure: Be transparent about AI usage
  • Data Protection: Safeguard institutional data
  • Content Responsibility: Verify AI outputs
  • Honor Code: Maintain academic integrity

Privacy

Support & Contacts