Microsoft and OpenAI disable Browse with Bing - the reason might surprise you
ChatGPT has lost its ability to offer you insights from web pages. What can we infer from this cryptic message from OpenAI?
The tool will pass the full parsed text to you from a website you asked it to go to. And why is that bad?
Let's suppose you asked it to visit a neo-Nazi website and the text it returns (factually correct in that it is the text from the website) is somewhat less than politically correct. Now let's suppose you do a screengrab of the text and produce a news article about how racist ChatGPT is.
The text displayed isn't being used as training data, it is being displayed because the user literally instructed the persona to display it.
So, what is a company to do? In this case, they simply decided to shut off the features that are being used to mislead folks. No effect on the tool's reasoning, logic or ethics controls, its learning, or its future generative outputs.
All because the loudest voice in the room doesn't want unfiltered propaganda being shown on screen? Or said differently, all because they cannot ensure the "right" propaganda is being shown on the screen?