Claude - Frequently Asked Questions

Do I retain ownership of my data I upload to the Claude platform?

The Anthropic Claude terms of service that govern the Syracuse University instance of Claude state that the user retains all ownership of any inputs to the Claude platform and is granted ownership over any outputs that Claude creates.

What classifications of data may be uploaded to the Claude platform when I logged in with my Syracuse University credentials to approved AI tools?

All classifications of university data can be uploaded and used when a user is logged in with university credentials (NetID).

How long does the Claude platform retain items like chat history, projects, artifacts, & items uploaded?

Currently Claude retains all data for 2 years. When you delete a chat or project, it is no longer visible to you and will no longer be part of your personalized memories in the platform. All deleted items are retained in the platform for the duration of the retention period independent of visibility.

What level of access does SU have to my AI usage in terms of data, prompts, and information input into the AI platforms?

As with all computing systems, data stored within Syracuse University information systems is inherently accessible to authorized IT staff in order to support, secure, maintain, upgrade, troubleshoot, and back-up those systems. 

At Syracuse, this access is strictly governed by the Information Technology Resources Acceptable Use policy (https://policies.syr.edu/policies/free-speech/information-technology-resources-acceptable-use-policy/) and the Information Security Framework and is exercised in accordance with departmental guidelines and best practices.

Does the Claude platform train on my data?

By default the data that Syracuse University users upload and the chat sessions they create are not used to train any of the Claude platform AI models. If you explicitly report feedback or bugs to Claude, then Claude may use your submission to train AI models.

The Claude platform has personalization features that are configurable by each user that allow for the platform to remember a user and use prior chat sessions and projects when generating new outputs. To change any memory or other personalization settings please visit the Claude Settings page. More information can be found at: Understanding Claude's Personalization Features | Anthropic Help Center

Are Incognito chats saved?

Incognito chats are temporary conversations that aren't saved to your chat history or to your personalized Claude memory. The chat session is still retained on the Claude platform and follows the retention policy of all other chats and projects.

How can I get access to Claude Code?

Claude Code is a way to consume Claude models via API access and the Claude Code terminal. The Claude Enterprise License does not include Claude Code, but users can purchase access via a credit card. Learn More.

Can I access Claude while traveling to other countries?

Anthropic's restricts access from certain countries. Attempting to access Claude from a prohibited country may result in your Syracuse University Claude account being banned by Anthropic. If your account is banned, please see: https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/488210484/Claude+-+Frequently+Asked+Questions#My-SU-Claude-account-was-banned.-Can-you-help%3F

For a complete list of countries where Claude is available, visit: Anthropic | Supported Countries

My SU Claude account was banned. Can you help?

If your account was banned by Anthropic, you'll need to submit an appeal directly to them through their support system:

Appeal process: https://support.claude.com/en/articles/8241253-safeguards-warnings-and-appeals

When submitting your appeal:

• Be honest and accurate about what happened

• Review Anthropic's public Terms of Service and Usage Policy to understand what may have triggered the ban

• Anthropic only allows access from some countries, see: Anthropic | Supported Countries

• Explain your intended use case clearly

Note that appeals are reviewed by Anthropic's Trust & Safety team, and approval is not guaranteed. While Syracuse University has an Enterprise agreement with Anthropic, we cannot influence their account ban decisions or appeal outcomes.

What is the difference between Claude Haiku, Sonnet, and Opus?

These are three different versions of Claude, each optimized for different needs:

Claude Haiku - Fastest and most efficient. Best for quick, straightforward tasks.
Claude Sonnet - Balanced performance and speed. Recommended for everyday use and most general tasks.
Claude Opus - Most capable and intelligent. Best for complex tasks requiring advanced reasoning and analysis.

Think of them as small, medium, and large options - choose based on the complexity of your task.

Note: Larger models (Opus and Sonnet) consume more of your usage limits than Haiku. Consider using Haiku for simpler tasks to preserve your capacity for more complex work using Sonnet or Opus.

How do I reset or extend Claude limits or access Claude Code?

In general, Claude limits reset through the day. Requests submitted through the Claude interface asking for usage limit increases has been disabled and requests will not be granted outside of the processes outlined below.

Claude Code and Claude API require a premium seat, more information found here: https://answers.atlassian.syr.edu/wiki/x/GQA_I

It is recommended you follow the best practices when using Claude or other AI systems to not exhaust limits unnecessarily. Some additional information from Claude: https://support.claude.com/en/articles/11647753-understanding-usage-and-length-limits

Staff and Faculty:

A Premium seat provides:

1. Increased usage limits per month compared to standard access

2. Access to additional features including Claude Code and the Claude API

Regarding usage limits:

While Premium provides higher limits, it does still have usage caps. If you anticipate high usage that may exceed these limits, we can configure an overage billing option. This works as follows:

1. Once you reach your Premium usage limit, you can continue using Claude on a pay-as-you-go basis

2. Overage charges are calculated per-usage and billed to the payment method associated with your Premium seat

3. We can establish a spending cap to prevent unexpected large bills

If you expect to require this overage billing configuration, please let us know a spending cap you are comfortable with and we can set it up for your account. Otherwise, your Premium package request should meet your needs.

We would recommend you try using the Premium seat as-is first, and if you find yourself hitting the limits often and are open to spending more for usage, please let us know by emailing aihelp@syr.edu.

Syracuse Faculty and Staff can use university funds to purchase the Premium license. These funds will most likely originate from your department and may require approval from a department head. To start the process, please request the the Claude Premium Access Package.

Students:

We do not currently provide premium seats for students. If you are a student worker and/or a department has authorized to pay for a premium seat for you, please have your supervisor and/or the department head reach out to us to set it up. This is subject to change, but the premium seat would be charged to the department at $2400/year, pro-rated from September 16th

I do not have approval for funding or do not wish to pay extra:

If you anticipate needing a higher usage limit and do not wish to pay for a Claude premium seat or credits, we would recommend you explore the http://mentor.ai.syr.edu platform to see if it can meet your needs. The mentorAI platform includes the Anthropic/Claude models. Usage limits in the mentorAI platform are higher and may offer tools Claude does not offer (screen sharing, etc)

What are some best practices when using Claude?

Claude offers several different models. Different models consume allocation at different “rates”.

For example, “Opus” tier model uses much more of your allocation per prompt than the “Sonnet” model. Identify which model you need for your task. If the task is routine (e.g., summarizing text, minor edits), pick the lighter model. If it’s complex (e.g., large architectural reasoning, research summary), use the heavier model - but be aware a heavier model will use a larger part of your usage limit. More information about Claude models in the above F.A.Q.

For Coding Tasks

• Provide complete context about your coding environment in your initial message.

• Include entire relevant code snippets in one message for reviews or debugging.

For Writing Assistance

• Outline requirements, target audience, and key points comprehensively.

• Send entire texts for editing in one message rather than breaking them up.

For Research and Analysis

• Clearly define your research question and focus areas initially.

• Provide all relevant data in a single, well-structured message.

By following these best practices, you can make the most efficient use of your Claude plan's message allocation.

Why can't Claude access the internet when running code?

When Claude executes code through its computer tools (bash, Python, R, etc.), network access is disabled. This is an intentional security configuration for Syracuse University's Claude Enterprise deployment.

Why is this restriction in place?

Claude's code execution environment has access to any data you bring into your conversation—whether through file uploads or connected services like Microsoft 365. Enabling network access would create a risk where code could potentially send sensitive University data to external endpoints. By keeping network access disabled, we ensure that any code Claude runs remains sandboxed and cannot transmit data outside the secure environment.

What can I do instead?

• Download data first: If you need to analyze data from an external source, download it to your computer and then upload it to Claude.

• Use Claude to write code: Ask Claude to help you write scripts that you can then run in your own local environment (RStudio, Jupyter, VS Code, etc.) where you have full control over network access.

• Web search is still available: Claude can still search the web and retrieve information conversationally—this restriction only applies to code execution.

Will this change in the future?

We continue to evaluate our security posture as Anthropic releases new features and controls. If a secure method for allowing limited network egress becomes available, we will consider enabling it.