Using CANChat

• CANChat uses an open-source user interface (OpenWebUI) licensed under the Massachusetts Institute of Technology.
• CANChat leverages large language models (LLMs) from third-party providers. You can switch between them depending on your task.

We encourage you to explore the CANChat user interface and take advantage of its many innovative functions and features.

• Bring your own data: Work with your own documents through CANChat’s upload function.
• Web searches: Use CANChat’s web search capabilities to quickly retrieve information from websites.
• Chat history: Review prior conversations with CANChat to resume discussions or revisit information.
• Chat archive: Use CANChat’s archival feature to securely store and reference past chats.
• Personalization: Tailor your CANChat experience with customizable settings to meet your preferences.
• Multi-model support: Benefit from various AI models in CANChat for diverse functionalities and expertise.
• Mobile accessibility: Connect with CANChat on-the-go with a mobile-friendly interface.
• Conversation tagging: Organize your discussions by tagging conversations, making them easy to sort and find.
• Language selection: Engage with CANChat in your preferred language.
• Input support: Interact with CANChat using voice commands for hands-free communication and convenience.

The following are acceptable uses of CANChat:

• Summarization: Quickly summarize documents, articles, policies, research papers, reports, meeting notes and more.
• Text translation: Draft translations of documents, presentations, reports and emails. Reminder that final translations always need to be reviewed by professional translators.
• Content creation: Be more creative with writing tasks by asking for suggestions or outlines. Note that if you use AI-generated content in final products, you need to follow the appropriate guidance to identify it. To learn more, visit: Guide on the use of generative artificial intelligence.
• Content editing: Adjust the tone, change the formality or reduce the length of content for more effective communication.
• Research and analysis: Researchers and analysts can save a lot of valuable time, especially when dealing with time-sensitive requests or a large number of documents.
• Meeting preparation: Generate agendas, talking points and summaries from transcripts and facilitate efficient meeting preparation and follow-up.

You are prohibited from:

• using departmental information with a classification higher than Unclassified
• using any personal, private or sensitive information
• engaging in unlawful or unethical practices
• generating harmful, offensive or malicious content
• attempting to bypass departmental security measures or exploit vulnerabilities

As a user, you must understand and be aware that generative AI tools have certain limitations. It is important to validate the outputs for accuracy, clarity, completeness and relevance.

Therefore, when using CANChat for work purposes, GC workers must:

• recognize that the effectiveness of AI models is influenced by the timeliness and age of their training data
• be aware of potential ethical concerns and ensure alignment with your organization’s values and goals
• understand that biased or incomplete data can lead to biased or inaccurate outcomes
• clearly identify when content is generated with assistance from generative AI tools; this helps maintain accountability and facilitates understanding of the AI-generated results by stakeholders, promoting trust and mitigating potential risks
• be aware of legal requirements, privacy regulations and intellectual property implications when working with SSC generative AI
• understand the potential risks associated with generative AI, such as the creation of misleading content or the amplification of biases

You agree:

• that all outputs generated by CANChat will be validated by a subject-matter expert prior to inclusion in a professional setting
• not to submit data or information above the Unclassified level of sensitivity
• not to use any personal, private or sensitive information
• not to use code generated by generative AI tools in a development or production setting, as it may violate copyright or intellectual property rights or result in security vulnerabilities
• not to engage in unlawful or unethical practices
• not to generate harmful, offensive or malicious content
• not to attempt to circumvent departmental security measures or attempt to exploit vulnerabilities
• that there is no expectation of solution support, and you understand that the product can be decommissioned without notice to the user