A programming framework for agentic AI
Update TRANSPARENCY_FAQS.md (#618)
Jack Gerritsbbd8ead43eaa
diff --git a/TRANSPARENCY_FAQS.md b/TRANSPARENCY_FAQS.mdindex c16f2685e..3d6489f1e 100644--- a/TRANSPARENCY_FAQS.md+++ b/TRANSPARENCY_FAQS.md@@ -54,7 +54,7 @@ Additionally, AutoGen’s multi-agent framework may amplify or introduce additio- Code execution: AutoGen recommends using docker containers so that code execution can happen in a safer manner. Users can use function call instead of free-form code to execute pre-defined functions only. That helps increase the reliability and safety. Users can customize the code execution environment to tailor to their requirements.- Human involvement: AutoGen prioritizes human involvement in multi agent conversation. The overseers can step in to give feedback to agents and steer them in the correct direction. In all examples, users confirm code before it is executed.- Agent modularity: Modularity allows agents to have different levels of information access. Additional agents can assume roles that help keep other agents in check. For example, one can easily add a dedicated agent to play the role of safeguard.-- LLMs: Users can choose the LLM that is optimized for responsible use. The default LLM in all examples is GPT-4o which inherits the existing RAI mechanisms and filters from the LLM provider. We encourage developers to review [OpenAI’s Usage policies](https://openai.com/policies/usage-policies) and [Azure OpenAI’s Code of Conduct](https://learn.microsoft.com/en-us/legal/cognitive-services/openai/code-of-conduct) when using GPT-4o.+- LLMs: Users can choose the LLM that is optimized for responsible use. The default LLM in all examples is GPT-4o which inherits the existing RAI mechanisms and filters from the LLM provider. We encourage developers to review [OpenAI’s Usage policies](https://openai.com/policies/usage-policies) and [Azure OpenAI’s Code of Conduct](https://learn.microsoft.com/en-us/legal/cognitive-services/openai/code-of-conduct) when using GPT-4o. We encourage developers experimenting with agents to add content moderation and/or use safety metaprompts when using agents, like they would do when using LLMs.