The answers may sound authoritative, but it’s well documented that Googles AI search results give out millions of inaccurate results an hour. They don’t like to talk about that. It’s an uncomfortable truth that runs counter to the narrative that AI is the second coming.
It’s also an existential threat to AI companies non-existent revenue streams as a recent German court ruling has laid out the legal framework for AI companies being held accountable for the consequences of erroneous results. In short, AI results are not protected speech as they are new algorithmic creations, which also bars them from the protections offered by defining them as simple summaries.
However, the German court held that this safeguard no longer applies when search engines incorporate generative AI systems. According to its reasoning, this technology is capable of producing nonexistent claims based on multiple sources and, consequently, the companies responsible for operating it must assume liability for the resulting content.
This extends far beyond the realms of search results on your phone and desktop, and, as I experienced last night, right into smart devices in my home. We’re under a heat warning here right now, and as I prepared for bed last night, I asked Google, “What’s the temperature outside?”. I knew the temperatures were supposed to drop overnight, and I wanted to know if it was a good time to open the windows.
Google said it was 23 C. Shit. I opened my phone to see the hourly forecast, only to have the widget on my home screen inform me it was 17 C. I asked Google again. And the conversation went a little something like this:
Me: Hey Google, what’s the temperature outside?
Google: It’s 17 C outside.
Me: But you just said it’s 23?
Google: I’m sorry. I got that wrong. It is 17 C outside.
— I wait a minute while I look up the temperature on the Weather Network (it’s 17 C).
Me: Hey Google, what’s the temperature outside?
Google: It’s 23 C outside.
Me: But you just said it’s 17.
Google: I’m sorry, I got that wrong. It is in fact, 17 C outside.
Me: So which is it? 17 or 23?
Google: It is 23 C outside. My previous information was incorrect.
I throw my hands up in frustration and ask Google (oh, the irony) to turn out the lights.
This can’t be the future of search. The level of inaccuracy and our distrust of search results, coupled with Google Geminis apparent inability to perform the functions that Google Assistant used to accomplish easily, has led to us considering completely removing all of our smart home devices.
Not to mention that Google often waffles on for minutes on end, answering something that you didn’t ask, and being so wrapped up in its own bullshit that it doesn’t hear you telling it to stop until you’re standing over the device yelling threats about sledgehammers while the veins pop on your neck and forehead.
This is not convenience. This is not a better life, or an easier way of doing things. This is the opposite.
End users on voice, desktop or mobile are doing more work to verify the results than a simple search used to require, and it is coupled with a minor anxiety (as if we needed another one) that builds over time because it cannot be trusted.
All this in service of a future that is being forced upon us.
This is not the Star Trek future I was promised. Like Riker, I wanted to speak to the ceiling and receive the answers I needed. I had that with Google Assistant.
It wasn’t perfect, but it was substantially better that this misinformation spewing, information gathering, slow torture device. Did the Romulans plan this?
Join The Conversation: