I've been thinking about #chatGPT again, and specifically Microsoft's plans to add the technology to Bing. I have concerns.

A chat like interface to search is clearly going to increase the volume of zero-click searches (searches where the user gets the answer without having to click through to a web result). We see this already with Google rich results, but with GPT based results there is unlikely to even be an specific identifiable source that can be cited for any result. ... /cont

Why go looking for a website to visit when the answer is already in front of you?

Many of the billions of pages that were crawled to train chatGPT are dependent on search traffic. Without that traffic website revenue dries up and the information ceases to be published. Whilst chatGPT will retain knowledge of what was previously published, it will never be able to update it's information from any publisher who has shut up shop... /cont

That leaves us with the odd situation where chatGPT powered search could lead to the deterioration of its own data sources.

Presumable a chatGPT bing feature will be using the openAI crawl rather than bingbots own crawl. As far as I have been able to find, there is no practical way yet to block chatGPT from crawling and absorbing your own content. ../cont

This means that a publisher with unique content has no means to ensure it is not absorbed en masse by an entity that could then cannibalise their traffic. In terms of copyright, I am sure OpenAI will rely on the argument that any reproduction of content is transformative and therefore not protected. That won't be much relief to the publisher if chatgpt bing takes off.