Microsoft knows you love tricking its AI chatbots into doing weird stuff and it’s designing ‘prompt shields’ to stop you
==========

Microsoft is developing new safety features for Azure AI Studio to prevent people from tricking AI chatbots. The features include 'prompt shields' to detect and block deliberate attempts to make an AI model behave in an unintended way. Microsoft is also addressing 'indirect prompt injections' where hackers insert malicious instructions into the data a model is trained on. The company aims to boost trust in its generative AI tools and is rolling out a feature that alerts users when a model generates erroneous responses. Microsoft is OpenAI's largest investor and is committed to deploying AI safely.

#Microsoft #AiChatbots #AzureAiStudio #PromptShields #GenerativeAi #Openai

https://fortune.com/2024/03/29/microsoft-openai-chatbots-prompt-shields-safety-features-user-tricks/