Microsoft Unveils Controversial Feature That Allows Companies To Hand Over Sensitive Data In Return For Better AI Results

Computational giant Microsoft is turning plenty of heads when it unveiled a new feature that’s being called out as controversial for obvious reasons.

The software giant says it wants to take in corporate data in exchange for AI-generated results that are better. This would be done through the firm’s Azure OpenAI Service which yields better results when you put forward questions to an AI chatbot.

Remember, OpenAI is one firm that’s linked to the popular ChatGPT tool and we know very well how keen they are in terms of investing in the corporation. In recent times, the investment has gone up to billions for the thorough development and now, the tool is being hosted across the Azure server.

After that, we witnessed Redmond being incorporated into AI models that are behind the whole thing. It’s the latest version of ChatGPT-4 and becoming a part of the entire range of goods and services is a huge thing.

So many major concerns linked to privacy and security have been brought up since the launch and that is linked to the launch of ChatGPT, ever since it was unveiled out in the open.

Today, it’s getting all sorts of data from various sources while undergoing training and it keeps on collecting data. But tech giant Microsoft might seem to be going in another direction.

Their new feature of handing over data in return for better results might soon turn into a feature that is highly requested.

The concept has to do with a simple notion- the more the data is up for grabs, the better the quality of results. Such types of systems work by fetching in more data of relevance that’s internal to a company so they can best complete anyone’s request for help.

Then we saw Microsoft mention how this combined with Cognitive Search from Microsoft Azure would help to see which information needs to be dug out from any kind of data source that is linked to user input and gives conversation history.

Such information then gets readded as prompts to the AI model owned by OpenAI and it would then dig out the right type of information related to the prompt.

Furthermore, experts claim such features give rise to no need for unnecessary training of AI models.

H/T: Theregister

Read next: What’s the Best AI Chatbot? This Study Might Hold the Answer
Previous Post Next Post