
"We're working on blocking them from being indexed now"
Why didn't you work on blocking them before you made this "tool" available ?
Because you always wait for the backlash before doing something that might impede your ability to slurp more data.
Google's Bard chatbot is currently being re-educated to better understand privacy. In July, Bard gained the ability to share conversations with other people using a unique public link. Unfortunately, Google Search has indexed those shared links, making them more widely available and discoverable than Bard patrons might expect …
It can be both.
In this case, it looks like a combination of a systemic desire to harvest as much data as possible – which includes publishing much of it, because then you get additional data about who wants it – and a design failure that arose in part from the culture produced by that systemic problem.
Honestly, though, it's already pretty good.
I wouldn't let this small set back get your train off the rails, Google UK.
On the other hand, indexing chat results? I hope it's fixed already with all your focus on privacy within the main Google Bard app.
Do you want to write some fantasy?
https://bard.google.com/chat
The example questions are fine to learn how to address the (currently) text based AI. It seems to know about ALL the freeware, including tools to query the whole USB bus system and see when things get plugged in... :)
Never mind the privacy issues, this is nightmare fuel to the singularity folks... Search engine indexes AI chats. The bots get trained on those indexes, thus learning from each and from their own mistakes in some kind of bizarrely exponentially increasing positive feedback loop that has no ceiling. Something, something, something. Humanity is wiped out a nanosecond after someone realises there's a real problem.