If you think that chatgpt can replace Google, I recommend you to think again

The current discourse in technology forums is portraying a narrative according to which chatgpt can be a replacement to Google. A fellow tech enthusiast, who is a popular podcaster and who I highly appreciate, Dror Globerman , has written a post about it (for the Hebrew readers) describing the situation as an earthquake, no less, that can bring Google down. Dror claimed, that the introduction of #chatgpt made #google service obsolete and inferior and that once chatgpt will know more, get more updated, be more accurate and international, then we all will need to block some time on our calendars to attend the virtual funeral ceremony of Google, may it rest in peace.

Now, regardless of the fact whether or not Google already has a similar or even superior technology to chatgpt that they, for the time being, choose not to deploy for one reason or another, I think that there is another angle one should consider before joining the crowd in the exclamations of joy – The King is Dead, Long Live the King!

In some contexts, chatgpt might be more friendly and intuitive to use. We are used to conversing, from the moment we learn to speak, in order to gather information. We also tend to be convinced when speaking to people whom we consider to be smarter or more knowledgeable than us. Charisma is a bonus to get us convinced even quicker. We ask them a question. They answer. We ask them follow-up questions, and they answer those too. Their answers make sense and we adopt the narrative.

Everyone knows that the knowledge that can be stored on an electronic device is far greater than the capacity of our human brain. Put an advanced Artificial Intelligence model on top of it, and you get a super smart digital “being” that some might even consider charismatic.

Forget for a second the no-feelings, no-consciousness – the obvious shortcomings of chatgpt that were ground to dust in the latest tech discourse. The fact is that some might tend to regard chatgpt as a “being”, a smart “being”, a single “being” that you converse with, that is a good source of information to learn from. It is like having a conversation with a university professor who is a matter expert. Even if you don’t completely agree with them at first, the effect of them being a professor, and you being a common folk takes its course. So, you try to challenge them, but they have some good answers. You try again to attack it, maybe from a different angle, and they have even better answers until you get convinced.

This is where the comparison between the shiny new chatgpt and the good ol’ Google kicks in. It is true that Google will match its results to the person who made the search based on their acquaintance with that person (and Google is well acquainted with all of us apparently). There’s a reason why one person who will search for ‘Jaguar’ will get top results on the modern luxury car, and another person that ran the exact same search will get results about the super predator cat.

But Google will return many results. Some will be overlapping in the information they provide; others will have different pieces of information that in some cases even contradict each other. And this is the part in our learning process that is so important which Google provides and that chatgpt does not. It gives us human beings all sorts of information about a matter and makes us flex that thinking muscle. We need to think and research to figure out for ourselves what is true and what is not. What source of information we can trust and which one we don’t? Chatgpt is getting that thinking muscle numb.

And this is a point of concern. Let me give you an example. I had some free time on my hand so I decided to run the following experiment to formulate my position on Che Guevara. Was he a non-violent freedom fighter, or was he a ruthless killer?

I started with chatgpt. To put things into context, I have asked it some questions about Bin Laden first. Chatgpt determined decisively that “Osama bin Laden was a Saudi Arabian terrorist”. He is not a freedom fighter by any means, chatgpt insisted. Bin-Laden consciously committed acts of terror by killing civilians and “Freedom fighters seek to bring about political or social change through non-violent means”, chatgpt claimed. It was time to bring Che into my conversation with it. When I asked about Che, chatgpt changed the level of decisiveness. First, it replied that Che was an “Argentine Marxist revolutionary who played a key role in the Cuban Revolution”. When I claimed that he too used violence however he’s being considered by many as a freedom fighter, chatgpt answered that “it is difficult to determine”. Chatgpt made it very clear – Bin-Ladin is a terrorist while Che Guevara is a revolutionist that might have been a bit violent here and there or might not. I then ask Google straight up: “Did Che Guevara kill civilians?”. The results Google returned did not romanticize Che. Some pointed out he was a revolutionist, but others determined that the dude was a cold-blooded killer.

We don’t like to sweat over things anymore, I get it. We take the car, even when it is a fifteen minutes walk. We don’t climb the stairs if there’s an elevator around. If it can be done by a machine, let the machine do it. “Just Do It” is a great slogan, but we tend to like it more when others can “Just Do It” for us. It is easier that way so what the hell right? Wrong! It gets dangerous when we let others think for us, or when we trust an AI oracle to provide us with answers that we accept with no questions asked.

Therefore, I claim that we should not replace Google with chatgpt in its current form. It is really cool, I agree. And it can also be a great tool to generate Python scripts based on natural language descriptions or to provide accurate information on factual non-disputable matters (like the statistics on the great basketball show that flu-stricken Jordan put up in the Bulls vs. Jazz game in 97 finals, for example). But when it comes to research, learning, and building a body of knowledge – do yourself and your fellow human beings a favor, do not let that thinking muscle of yours take a rest. Keep on flexing it. We depend on it as a species.