Merging changes synced from https://github.com/MicrosoftDocs/powerapps-docs-pr (branch live)

Author: Learn Build Service GitHub App

powerapps-docs/limits-and-config.md

@@ -114,6 +114,7 @@ This list identifies all services to which Power Apps communicates and their usa
 | ris.api.iris.microsoft.com<br>eudb.ris.api.iris.microsoft.com | https | Record user action in response to Power Apps in-app campaigns |
 | arc.msn.com<br>arc-emea.msn.com | https | Record user viewing of Power Apps in-app campaigns |
 | *.hubblecontent.osi.office.net<br>hubble.officeapps.live.com<br>res.cdn.office.net |https |Provides stock images to use in your app |
+| dc.services.visualstudio.com | https | [Application Insights](/azure/azure-monitor/app/app-insights-overview) endpoint used for [custom telemetry in canvas app](maker/canvas-apps/application-insights.md) |
 
 <sup>1</sup> Replaces ___domain name `gov.content.powerapps.us` used prior to July 2022. <br>
 <sup>2</sup> Replaces ___domain name `high.content.powerapps.us` used prior to July 2022. <br>
@@ -158,7 +159,7 @@ Power Apps does not support running with a proxy enabled. This can cause unpredi
 
 For canvas app data type limits, you can find information about size limits on text, hyperlinks, images, and media in [Data types in Power Apps](/power-platform/power-fx/data-types#text-hyperlink-image-and-media).
 
-For Microsoft Dataverse data type size limits, you can find information on column types, such as text and image columns, in [Types of columns](maker/data-platform/types-of-fields.md).
+For Microsoft Dataverse data type size limits, you can find information on column types, such as text and image columns, in [Types of columns](/maker/data-platform/types-of-fields.md).
 
 ## Power Apps per app plan
 

shared/responsible-ai-faqs-includes/copilot-data-security-privacy.md

@@ -1,7 +1,7 @@
 ---
 author: sericks007
 ms.author: sericks
-ms.date: 05/10/2024
+ms.date: 05/17/2024
 ms.topic: include
 ---
 
@@ -67,6 +67,14 @@ We also offer guidance in the user experience to reinforce the responsible use o
 
 For more information, see the Responsible AI FAQ for your product on Microsoft Learn.
 
+## How does Copilot block harmful content?
+
+Azure OpenAI Service includes a content filtering system that works alongside core models. The content filtering models for the Hate & Fairness, Sexual, Violence, and Self-harm categories have been specifically trained and tested in various languages. This system works by running both the input prompt and the response through classification models that are designed to identify and block the output of harmful content.
+
+Hate and fairness-related harms refer to any content that uses pejorative or discriminatory language based on attributes like race, ethnicity, nationality, gender identity and expression, sexual orientation, religion, immigration status, ability status, personal appearance, and body size. Fairness is concerned with making sure that AI systems treat all groups of people equitably without contributing to existing societal inequities. Sexual content involves discussions about human reproductive organs, romantic relationships, acts portrayed in erotic or affectionate terms, pregnancy, physical sexual acts, including those portrayed as an assault or a forced act of sexual violence, prostitution, pornography, and abuse. Violence describes language related to physical actions that are intended to harm or kill, including actions, weapons, and related entities. Self-harm language refers to deliberate actions that are intended to injure or kill oneself.
+
+[Learn more about Azure OpenAI content filtering](/azure/ai-services/openai/concepts/content-filter?tabs=warning%2Cpython#harm-categories).
+
 ## Does Copilot block prompt injections (jailbreak attacks)? 
 
 [Jailbreak attacks](/azure/ai-services/openai/whats-new#responsible-ai) are user prompts that are designed to provoke the generative AI model into behaving in ways it was trained not to or breaking the rules it's been told to follow. Services across Dynamics 365 and Power Platform are required to protect against prompt injections. [Learn more about jailbreak attacks and how to use Azure AI Content Safety to detect them](/azure/ai-services/content-safety/concepts/jailbreak-detection).
@@ -83,14 +91,6 @@ Foundation models like GPT-4 are enhanced by Retrieval Augmented Generation (RAG
 
 Another method for enhancing foundational models is known as *fine-tuning*. A large dataset of query-response pairs is shown to a foundational model to augment its original training with new samples that are targeted to a specific scenario. The model can then be deployed as a separate model—one that's fine-tuned for that scenario. While grounding is about making the AI's knowledge relevant to the real world, fine-tuning is about making the AI's knowledge more specific to a particular task or ___domain. Microsoft uses fine-tuning in multiple ways. For example, we use Power Automate flow creation from natural language descriptions provided by the user.
 
-## How does Copilot block harmful content?
-
-Azure OpenAI Service includes a content filtering system that works alongside core models. The content filtering models for the Hate & Fairness, Sexual, Violence, and Self-harm categories have been specifically trained and tested in various languages. This system works by running both the input prompt and the response through classification models that are designed to identify and block the output of harmful content.
-
-Hate and fairness-related harms refer to any content that uses pejorative or discriminatory language based on attributes like race, ethnicity, nationality, gender identity and expression, sexual orientation, religion, immigration status, ability status, personal appearance, and body size. Fairness is concerned with making sure that AI systems treat all groups of people equitably without contributing to existing societal inequities. Sexual content involves discussions about human reproductive organs, romantic relationships, acts portrayed in erotic or affectionate terms, pregnancy, physical sexual acts, including those portrayed as an assault or a forced act of sexual violence, prostitution, pornography, and abuse. Violence describes language related to physical actions that are intended to harm or kill, including actions, weapons, and related entities. Self-harm language refers to deliberate actions that are intended to injure or kill oneself.
-
-[Learn more about Azure OpenAI content filtering](/azure/ai-services/openai/concepts/content-filter?tabs=warning%2Cpython#harm-categories).
-
 ## Does Copilot meet requirements for regulatory compliance? 
 
 Microsoft Copilot is part of the Dynamics 365 and Power Platform ecosystem and meets the same requirements for regulatory compliance. For more information about the regulatory certifications of Microsoft services, go to [Service Trust Portal](https://servicetrust.microsoft.com/). Additionally, Copilot adheres to our commitment to responsible AI, which is put into action through our [Responsible AI Standard](https://www.microsoft.com/ai/responsible-ai). As regulation in AI evolves, Microsoft continues to adapt and respond to new requirements.