𝗝𝘂𝗻𝗲 𝟮𝟬𝟮𝟱 𝗨𝗽𝗱𝗮𝘁𝗲: 𝗣𝗜𝗜 𝗗𝗲𝘁𝗲𝗰𝘁𝗶𝗼𝗻 𝗖𝗼𝗻𝘁𝗲𝗻𝘁 𝗙𝗶𝗹𝘁𝗲𝗿𝗶𝗻𝗴

When using Azure OpenAI models (like GPT-4, GPT-3.5, o3-mini), 𝗱𝗲𝗳𝗮𝘂𝗹𝘁 𝘀𝗮𝗳𝗲𝘁𝘆 𝘀𝗲𝘁𝘁𝗶𝗻𝗴𝘀 are already applied to protect users and ensure responsible use. These include:

• 𝗖𝗼𝗻𝘁𝗲𝗻𝘁 𝗙𝗶𝗹𝘁𝗲𝗿𝗶𝗻𝗴 – blocks harmful or unsafe content
• 𝗕𝗹𝗼𝗰𝗸𝗹𝗶𝘀𝘁𝘀 – prevent certain words or topics
• 𝗣𝗿𝗼𝗺𝗽𝘁 𝗧𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻 – rewrites risky input
• 𝗖𝗼𝗻𝘁𝗲𝗻𝘁 𝗖𝗿𝗲𝗱𝗲𝗻𝘁𝗶𝗮𝗹𝘀 – tracks generated content

However, 𝗣𝗜𝗜 (𝗣𝗲𝗿𝘀𝗼𝗻𝗮𝗹𝗹𝘆 𝗜𝗱𝗲𝗻𝘁𝗶𝗳𝗶𝗮𝗯𝗹𝗲 𝗜𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻) 𝗱𝗲𝘁𝗲𝗰𝘁𝗶𝗼𝗻 is 𝗻𝗼𝘁 𝗯𝘂𝗶𝗹𝘁-𝗶𝗻 directly to these models. You need to 𝗮𝗱𝗱 𝘁𝗵𝗶𝘀 𝗺𝗮𝗻𝘂𝗮𝗹𝗹𝘆 using 𝗔𝘇𝘂𝗿𝗲 𝗔𝗜 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 or 𝗔𝘇𝘂𝗿𝗲 𝗖𝗼𝗻𝘁𝗲𝗻𝘁 𝗦𝗮𝗳𝗲𝘁𝘆 𝗔𝗣𝗜𝘀.

𝗣𝗜𝗜 𝗗𝗲𝘁𝗲𝗰𝘁𝗶𝗼𝗻:

PII detection helps identify and filter private details like:

• Names
• Phone numbers
• Email addresses
• Addresses
• Government ID numbers

This protects user privacy and keeps AI output safe and compliant.

𝗘𝘅𝗮𝗺𝗽𝗹𝗲:

User input:
"Hi, I'm Tanuja. My phone number is 7283386677."

PII detection will convert it to:
"Hi, I'm [NAME]. My phone number is [PHONE_NUMBER]."

You can safely pass this version to Azure OpenAI for processing.

𝗛𝗮𝘃𝗲 𝘆𝗼𝘂 𝗶𝗺𝗽𝗹𝗲𝗺𝗲𝗻𝘁𝗲𝗱 𝗣𝗜𝗜 𝗱𝗲𝘁𝗲𝗰𝘁𝗶𝗼𝗻 𝗶𝗻 𝘆𝗼𝘂𝗿 𝗔𝗜 𝗮𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 𝘆𝗲𝘁?
𝗜𝗳 𝘆𝗲𝘀, 𝘄𝗵𝗶𝗰𝗵 𝘀𝗲𝗿𝘃𝗶𝗰𝗲 𝗼𝗿 𝗺𝗲𝘁𝗵𝗼𝗱 𝗱𝗶𝗱 𝘆𝗼𝘂 𝗰𝗵𝗼𝗼𝘀𝗲?