Artificial Intelligence (AI)

Introduction

The rapid development of Artificial Intelligence (AI) systems brings significant opportunities,
alongside technological, ethical, legal, and academic risks.

This document establishes guiding principles for the responsible, cautious, and secure use of AI tools
across all areas of activity at the Technion. Its purpose is to ensure appropriate academic standards,
protect personal information, and enable informed and risk-aware use of AI tools,
in accordance with institutional procedures, privacy protection laws, and information security regulations.

Target Audience

These guidelines apply to the entire Technion community – academic and administrative staff,
teaching staff, students, and visitors operating within the Technion.

Key Risks in Using AI Tools

  • Information Security Risks: Data leaks, exposure of Technion systems and databases to attacks, and bypassing existing authorization mechanisms.
  • Privacy Risks: Leakage of personal and/or confidential information or unauthorized use of personal data. Violations of privacy protection laws may result in criminal sanctions and fines.
  • Organizational Risks: Uncontrolled reliance on AI outputs may lead to biased, incorrect, or discriminatory decisions.
  • Exposure of Sensitive Organizational Information: Uploading information to AI tools may result in long-term exposure of such data.
  • Violation of Intellectual Property Rights.
  • Breach of Confidentiality Obligations in discussions involving personal matters.

Responsibility

Artificial intelligence tools are assistive tools for users.
The outputs generated and their consequences remain the sole responsibility of the user.

Organizational AI Tools at the Technion

The Technion provides organizational AI tools acquired under institutional agreements.
These tools operate within dedicated organizational environments that include advanced security
and protection mechanisms.

Currently available tools include Organizational Copilot and Organizational ChatGPT.
These tools are available for purchase through CIS Software and Services Shop.

When handling personal and/or confidential information, only organizational tools must be used.

Golden Rules for Using AI Tools at the Technion

1. Prefer Organizational Tools

Use AI tools that are approved and managed by the Technion rather than private or free external tools.

2. Provide Only the Minimum Necessary Information

Before entering information, ask yourself: Is it truly necessary to provide all of this data?

3. Do Not Enter Personal or Confidential Information into External Tools

Doing so may violate privacy laws and expose both the user and the Technion to legal and regulatory consequences.
Personal data should be anonymized when included in organizational files.

4. Always Validate AI Outputs

AI systems may generate incorrect or misleading content. Always review, think critically,
and verify information against reliable human or authoritative sources.

5. Do Not Connect AI Agents to Core Systems Without Approval

Connecting AI agents without authorization may introduce significant security risks.

6. New AI Tools Require Approval

Before adopting a new AI tool, confirm that no suitable organizational solution exists
and obtain approval from the Computing and Information Systems Division.

7. Review Contracts and Agreements

Verify where data is stored and whether it is used for training general models before implementing
systems that incorporate AI functionality.

8. Responsible Use in Teaching, Research, Administration, and Personal Discussions

  • Teaching: Ensure compliance with intellectual property rights and protection of personal data.
  • Research: Do not upload confidential research data to external cloud services without approval.
  • Administration: AI-generated outputs must be reviewed before use.
  • Personal Discussions: Full confidentiality obligations apply in committee discussions involving personal matters.

9. When in Doubt – Ask

If you are uncertain or concerned about potential exposure, contact the Computing and Information Systems Division.

Frequently Asked Questions

General
  • Can undergraduate students purchase through the Technion Software Shop?
    Only Technion budget account holders can make purchases through the Software Shop. Licenses may also be purchased for students using a Technion budget account.
  • Must a faculty member purchase licenses for students in their research group? Can students purchase independently and be reimbursed?
    Purchases cannot be made using credit cards. All purchases must be made through Technion budget accounts. The Software Shop is not a standard online store with credit card processing, but a system integrated with SAP for charging Technion budget accounts. A license can be purchased from a unit’s budget even if it is assigned to a student.
  • Is there an expected timeline for adding Claude or Gemini licenses?
    Regarding Gemini – an agreement with MACHBA exists, and the matter will be discussed and decided together with management as part of Gmail license procurement. Other tools depend on future MACHBA agreements.
Copilot
  • Does every Technion student or employee have access to a free Copilot account?
    Yes. Anyone with a Technion Microsoft 365 account has access.
  • Can agents developed under a paid account be shared with free accounts?
    Agents developed within the organization can be shared. Agents available as part of the additional paid license cannot be shared.
  • Is there support for VS Code? How is onboarding handled?
    Support is provided through GitHub Copilot, for teaching purposes only.
  • Can Claude models be selected? Which models are supported?
    Models are available within GitHub as part of the free educational licensing.
ChatGPT
  • What should be done if an API is required that is not supported under the Technion license?
    A review must be conducted with Information Security, MACHBA, and OpenAI support prior to use.
  • Can agents developed within the organization be shared?
    Agents available as part of the additional paid license cannot be shared.
  • Is there support for VS Code? How is onboarding handled?
    Support is provided through GitHub Copilot, for teaching purposes only.
  • Can Claude models be selected? Which models are supported?
    Models are available within GitHub as part of the free educational licensing.

Relevant Links:
GitHub Education – Students
GitHub Education – Teachers

Contact and Reporting

In case of suspected exposure, leakage, or unauthorized use of personal and/or confidential information:

Thank you for your cooperation.

Related Topics :