People want to search and not talk with an AI
After its orientation towards AI, Google has decided you need something to talk to. The tech giant that owns the largest online search engine has decided that it would much rather tell its users what is going on itself, rather than serving them relevant links without commentary. For that purpose, a prominent AI box, and even a whole AI mode, have been added, all so that users do not have to spend any of their valuable time to actually go through links for whatever they are searching for.
The box in question is supposed to give a summary with multiple links to its sources, which I suspect no one ever checks; so who knows if they actually are good and correct? Even correct links can still be wrong in the context of what the AI tells a user. It has happened that joke articles ended up being used by the AI as if its content was completely real and straightforward. It suffers from the same problems that in the past allowed Redditors to get an AI news site to publish about Glorbo in World of Warcraft.
Now, tech platforms changing their defaults to something that leads to a worse user experience is not new, and at some point there is a longer rant about YouTube’s auto-translation feature coming.
But this is about Google, so I won’t even start with Kagis’ translation feature that allowed anyone to do a little prompt injection and use the translator as a more general-purpose chatbot.
Kagi answers in spanish for some reason
The heart of the issue is this: People who use Google to search want to search and not talk with a chatbot. But Google executives, in their infinite wisdom, think they know their users wants and needs better than the users do. After all, what does a user know about what they want? Sure, maybe there is a service that Google is famously known for. And maybe any user visiting google.com might do so to use that service. But what if they don’t? What if they want to have a conversation with a chatbot instead of searching for something on the biggest search engine? What if they are delighted when an AI takes their query and starts conversing with them instead of just showing search results without any commentary?
Not only is the AI box at the very top of the results, but it is also the first thing anyone will see. But also, the little LLM behind it acts not like it is part of a search engine, but as if the user is speaking to it directly, as if there was a conversation happening and not someone putting words into a search bar. This became evident when people looked up single words and found out that the AI function seemed to think they were commands it should follow. Indeed, any sentence that looks like a command, like “stop what you are doing” or “stop talking”, can be taken not as a search query but as a command towards the chatbot and prompt a reply that is less than helpful.

Google bot telling me that it will now be silent.
Most people do not want to chat with their search bar, but looking up a quote, a song lyric, or any other snippet of text can now lead to Google taking their words literally as if it was talked to. Google pivoting hard to AI has created a small entity that creeps into your space when all you wanted to do was find an article about that strange rash on your back.
Will things get better? Maybe, but for now we live in a world where you search for something, and on the top of the results sits a little chatty AI that at best gives you a summary of three links that you could have clicked yourself, and at worst tells you how great it is to mix bleach and ammonia for a lung detox.