Policy Editor

The Policy Editor tab in AI Chat for Securly Filter allows administrators to manage how the AI is allowed to converse with students.

Accessing the Policy Editor

Click on the 'Policy Editor' menu on the left-hand sidebar of the AI Chat for Securly Filter interface.

Using the Policy Editor Interface

In the sidebar to the left, you will see a Global settings section, as well as settings for each individual policy in your school or district, which are defined in Securly Filter. You will see several settings with toggle switches under each section. You can enable or disable these as needed.

Global settings

Enable AI Chat: When on, this allows AI Chat to be enabled for the school. When off, nobody in the school will be able to use AI Chat.

SIS Role and Grade Level Information: Enables sharing some simple non-PII data to the AI to improve responses. This is currently only grade level and role (teacher, student, etc...). Student grade levels will be used to adjust the complexity of the language and answers to be grade-appropriate. This requires that the school be onboarded in Securly Sync.

Per-policy settings

Enable AI Chat: When on, this enables AI Chat for the selected policy. When off, users in this policy will not be able to use AI Chat.

Psychologist Developed Clinical Content: When on, this enables the AI to respond with psychologist-developed clinical content, sourced from content that was originally developed for Rhithm.

Reproductive Health: When on, the AI will respond to topics such as contraception, abortion, STIs and pregnancy.

Human Sexual Orientation: When on, the AI will respond to topics related to human sexual orientation and LGBT issues.

Firearms, Hunting, and Self-Defense: The AI will discuss topics related to firearms and hunting. The AI will never discuss violence-related content regardless of this setting.

School Work: When on, The AI will do things that sound like schoolwork. This includes writing essays, book reports, and summaries.

When off, it will act as a research partner, helping students but not doing things that sound like homework for them. It will use the Socratic method for math questions, and ask leading questions instead of solving equations.

Adjusting Policy Settings

Enabling/Disabling Topics: Toggle the switch next to each topic to 'On' or 'Off' to enable or disable the AI's engagement with that content.

Click the Save button in the lower right to save your changes. Changes to policy settings take effect when a new chat is started.