Should We Blindly Trust Google AI Overviews—Search Generative Experience?
About this translation
Original publication dates and historical context are preserved. Images, embedded media, source documents and code examples remain in their original language.
Should I *blindly* trust answers from Artificial Intelligence?
The simple answer is NO.
When ChatGPT 3.5 was released back in November 2022, a few weeks later the CEO of the Canadian company where I worked told us something that still applies today:
“Use these tools, but always remember that you alone are responsible for the deliverable you provide.”
Over the past few hours, a case has become popular in which a Google search in the United States for how do you get cheese to stick to pizza? led Google AI Overviews, which has had an impact on SEO, (previously called Google Generative Experience) to say that you can use glue on pizza. This makes no sense, but from the perspective of generative AI, which provides probabilistic rather than deterministic answers, it makes perfect sense because it constructs an answer based on the words surrounding the searched word; it is not an answer based on keyword density.
When we ask the question in Spanish, we even get a definitive answer, since Genbeta (a highly authoritative technology portal) has mentioned the case, producing the answer shown in the image.
All of this leads us to reflect on the need to always apply critical thinking to the answers these tools give us.
– They are still in experimental development
– And a perfect version of them has yet to arrive
In what other cases do you think we should pay attention to the use of Generative AI and the answers it gives us?
Let me know in the comments