News

How Azure OpenAI Can Help Users Parse Unhelpful Error Messages

• By Kurt Mackie
• 05/01/2023

Organizations can use Azure OpenAI to find solutions to vague error messages, such as those that contain "GUIDS, seemingly random numbers, and vague text."

Microsoft demonstrated the capability last week in a video by Microsoft Cloud Solution Architect Greg Beaumont. Beaumont showed several examples of how Azure OpenAI using the ChatGPT 3.5-Turbo model can take error messages and deliver possible answers, which get formatted into an ordered list in plain English.

Beaumont used a simple prompt in the Azure OpenAI Studio's "Chat Playground" to tell the artificial intelligence (AI) to act as a chatbot that solves error messages about SQL Server, SharePoint, Teams and Power BI.

Next, Beaumont used actual forum questions posted by people who had problems trying to figure out error messages into the Chat Playground to get the AI model's responses. He then compared the answers to the actual forum answers. Typically, ChatGPT's No. 1 proposed solution matched the solution that was given in the forum, per the demo.

One point made repeatedly by Beaumont is that getting answers through the Azure OpenAI service doesn't expose sensitive information externally. He stated that "if a user enters sensitive data that is within the error message, it will be contained in your Azure environment."

The company OpenAI has access to the ChatGPT prompts submitted by users, which has only recently been highlighted in the news. For instance, Samsung employees reportedly exposed proprietary information while using ChatGPT to fix some source-code problems, according to various media accounts, such as this Techradar.pro article.

The notion that Azure OpenAI maintains internal organizational permissions also was demonstrated in the recent Microsoft Mechanics video, "Can ChatGPT Work with Your Enterprise Data?"

Microsoft's Azure OpenAI allows users to "interact with your organization's private internal data, while respecting the information protection controls put in place," argued Jessica Hawk, corporate vice president for data and AI, in this announcement.

Hawk further argued that organizations wanting to optimally use AI should "modernize their data estate" using the Microsoft Intelligent Data Platform, which is Microsoft's umbrella term for its various databases, analytics and data governance tools.