Contents
I. Privacy Notice (Global)
Click here to download Global Privacy Notice.
II. Guidance & Guidelines for Responsible Use of AI
Internal Use Only
Purpose and Scope
The Learning Experience Platform (LXP) leverages AI to enhance learning and development for EY personnel and our clients. This document provides guidelines and considerations for the responsible use of AI within the LXP, ensuring adherence to EY’s information security and ethical standards.
User Prompts and Data Usage
Rights and Permissions: Ensure you have the required rights and permissions to use the data you provide as prompts or upload as files within the LXP. This includes adhering to local hosting restrictions and only using client or third-party data if contractually permitted.
Sensitive Data: Do not enter or upload any sensitive personal information or sensitive personal data, including but not limited to:
- Personal data revealing racial or ethnic origin, political opinions, religious or philosophical beliefs, or trade-union membership.
- Genetic data, biometric data (for identification), data concerning health, or data concerning a natural person’s sex life or sexual orientation.
- Personal data relating to criminal convictions and offences.
- Tax file numbers or social security numbers/national identification numbers.
- Government client data, including Controlled Unclassified Information (CUI), audit secrets, state secrets, trade secrets, or any data violating professional secrecy or confidentiality rules.
Always Evaluate the Accuracy and Validity of AI Outputs
Accuracy and Validity: The LXP’s AI outputs may be outdated or inaccurate. Always evaluate the information provided by the AI and its sources before using it.
Non-Unique Outputs: AI outputs may not be unique and could be identical to outputs produced for other users. Ensure all work products using AI outputs, especially generated coursework, follow all country and/or Service Line specific quality processes and guidelines.
EY-Specific Answers: The LXP’s AI capabilities have not been trained on EY-specific information and cannot provide EY-specific answers. It only contains public information up until October 2023.
Responsible Use
Compliance: Comply with the Global Data Protection and Confidentiality Policy and use the AI capabilities ethically and responsibly.
Access the following link for more information: Data Protection Guidelines – AI Systems
Microsoft Policies: The AI is supported by Microsoft Azure infrastructure and leverages Azure OpenAI. Use must be compliant with:
Identifying and Reporting Hallucinations
Identifying Hallucinations: AI hallucinations occur when the AI generates information that is incorrect, misleading, or nonsensical. To identify hallucinations:
- Cross-check AI-generated information with reliable sources.
- Look for inconsistencies or information that seems implausible.
- Verify facts and data points provided by the AI.
Reporting Hallucinations: If you identify a hallucination:
- Document the hallucination with specific details.
- Report hallucinations via the AI Builder Support Form. This form is also available via the LXP interface and should be used for all AI support and feedback requests.
- Provide context and any supporting evidence to help address the issue.
Service Line Requirements
Assurance Service Line: Visit EY Atlas for detailed guidance applicable to all EY professionals supporting or participating as a member of an Assurance engagement team.
Strategy and Transactions Service Line: Approval from the SaT Engagement Leader is required for using AI during a client engagement. Document and retain such approval as part of the engagement work papers.
Tax and Consulting: Follow the general guidelines provided in this document.
Government & Public Sector: Do not enter Government client data, including Controlled Unclassified Information (CUI), into any generative AI tool, including the LXP.
Additional Information
For further details, contact the Global Data Protection team for assistance.
III. Free Text Field Disclaimer
EY does not intentionally collect any sensitive personal data from you via the Tool. The Tool's intention is not to process such information. Accordingly, you should not enter any unnecessary personal information or any sensitive personal data (including government identifiers such as tax file numbers or social security numbers/national identification numbers), client confidential information (except client entity names), audit secrets, state secrets, trade secrets, or anything that would violate professional secrecy or confidentiality rules or that would be considered abusive/irrelevant in the Tool.
Users should review the accuracy and completeness of all output generated though the Tool. Users should not enter any unnecessary personal information or any sensitive personal data, client confidential information, state secrets, commercial secrets, or anything that would violate professional secrecy or confidentiality rules or that would be considered abusive/irrelevant.