In these times, Artificial Intelligence has become more omnipresent than God – sorry believers. It seems to be everywhere. Everyone talks about it. Not a day goes by without someone praising its “intelligence.”
The fact that there are machines and algorithms capable of constructing sentences with a certain meaning (sometimes) or making videos with decent quality for the (inexperienced) human eye is not worrying, what is worrying is that we believe that this is intelligence.
“In the age of AI and Big Data, the real danger is not that computers are smarter than us. It’s just that we believe they are,” wrote Gary Smith, an economics professor at Pomona College.
Is Artificial Intelligence really intelligent?
Artificial Intelligence has crept into our lives so quickly that it is destroying with the same speed the skepticism and suspicion that it previously generated in us. A study carried out at Harvard University discovered that we already trust the advice given by algorithms more than that of people, even if they are specialists.
Recently, researchers from Italy and Switzerland also found that, when Artificial Intelligence has information about us, it is 82% more persuasive than humans. And it’s not exactly good news. Why?
For several reasons.
First, because Artificial Intelligence cannot carry out inductive reasoning or capture the true meaning of words. Sometimes its answers make sense, but that doesn’t mean the machine is aware of what it is saying. A parrot may repeat or mix up the words it has heard, but does not really understand their meaning.
Second, because Artificial Intelligence makes mistakes – a lot. It makes blunders (persuasively, of course). And if we believe what it tells us at face value, we will also be doomed to error. Trusting an algorithm by putting aside our rationality puts us at risk of developing that same obtuse and limited “thinking.”
Are we more stupid than 50 years ago?
It is possible.
In 2004, researchers at the University of Oslo analyzed more than 500,000 recruits and observed that after a peak in conventional intelligence occurred in the 1990s, something known as the “Flynn effect,” then our level of intellect has been decreasing.
They weren’t the only ones. Psychologists from the University of Adelaide in South Australia presented the results of 20 years of research with children between the ages of 6 and 13 and found that there has also been a slight decrease in IQ (approximately 7 points from 1980 to the beginning of the millennium).
And no, it does not only happen with logical intelligence.
One of the recent reports from the Organization for Economic Cooperation and Development (OECD) revealed, after analyzing the 38 countries that comprise it, that between 4.9 and 27.7% of adults show a level of competence extremely basic in reading comprehension.
That means they are only able to read and understand short texts on familiar topics. And they can only answer questions about reading if they find the answers verbatim, which means they don’t really process what they read.
Therefore, perhaps Artificial Intelligence is not so intelligent. Or maybe we should rethink what we consider “intelligence.” Or maybe we have an even bigger problem: we have stopped thinking and acting with common sense, so that anything seems smart to us.
The less we think, the more amazing Artificial Intelligence will seem to us
It is undeniable that machines have infinitely greater memory than ours, more access to data, they make calculations faster and they do not get tired like we do. But no matter how smart they seem, they are not able to develop expectations, make predictions based on intuition or put themselves in the shoes of others. And many times that is the key to solving the problems that most concern and affect us.
Artificial Intelligence has no way of checking its deductions with “common sense”, simply because it does not have it. That is why human reasoning and, above all, good sense and creativity, are more necessary than ever.
Intelligence does not rely solely on establishing connections quickly based on patterns detected after analyzing millions of data. Intelligence consists of analyzing problems from different angles and seeking creative solutions that make sense and are adaptive.
If we reduce the concept of intelligence to the “thinking” of machines, we run the risk of forgetting what true intelligence is. As philosopher Slavoj Žižek wrote: “The real danger is not that people will confuse a chatbot with a real person, but that communicating with chatbots will make real people talk like chatbots: ignoring all the nuances and ironies, obsessively focusing only on what someone thinks it is saying.”
The real danger is not that machines continue to progress and help us face the new challenges that will come, but that we think that this is intelligence because we have forgotten what human genius really is.
References:
Salvi, F. (2024) On the Conversational Persuasiveness of Large Language Models: A Randomized Controlled Trial. In: Cornell University; 2403.14380.
Žižek, S. (2023) Artificial Idiocy. In: Project Syndicate.
Ball, P. (2020) The AI delusion: why humans trump machines. In: Prospect.
Logg, J. M. et. Al. (2019) Algorithm Appreciation: People Prefer Algorithmic To Human Judgment. Organizational Behavior and Human Decision Processes; 151: 90-103.
Smith, G. (2018) Beware the AI delusion. In: FastCompany.
Batini, F. et. Al. (2013) OCSE: Skills Outlook. In: OECD.
Sundet, J. M.; Barlaug, D. G. & Torjussen, T. M. (2004) The end of the Flynn effect?: A study of secular trends in mean intelligence test scores of Norwegian conscripts during half a century. Intelligence; 32(4): 349-362.
Nettelbec, T. & Wilson, C. (2004) The Flynn effect: Smarter not faster. Intelligence; 32(1): 85-93.
