AI is designed to perform tasks that have traditionally required human intervention, including those that require judgment and decision-making. AI systems can quickly and accurately analyze huge amounts of data, which often exceeds our capabilities. This offers great benefits, such as increased efficiency and the ability to delegate repetitive tasks to machines. But by helping us to this extent and freeing us from difficult tasks, does it not risk diminishing us?
Not such a new question
In assessing the impact of generative artificial intelligence on learning and cognitive skills, it is tempting to draw parallels with the concerns raised by technology journalist Nicholas Carr, author in Atlantic 2008 of a significant article: “Is Google making us stupid?“. Two years later, he developed his remarks in his book Shallowspublished in French under the title Is the internet making you stupid? (ed. Laffont, 2011).
Nicholas Carr explored how the internet, and in particular the search engine Google, has redefined our reading habits, eroding our attention spans and reducing our ability to engage in deep, sustained thinking. His observations about his increasing difficulty concentrating and reading long texts resonated with many people, who noticed similar changes in their own reading practices.
The issue has had a global impact, prompting widespread debate and reflection on how our growing dependence on the Internet could fundamentally change our intellect. In essence, “Is Google making us stupid?” was the catalyst for a global reassessment of the relationship between technology and the human brain.
Do your cognitive exercises
In July 2023. American magazine Forbes published a powerful article entitled “Why ChatGPT makes us less intelligent“. He highlights legitimate concerns: the growing tendency to rely on instant responses, the decline of memory retention, the erosion of writing and communication skills.
Author and visionary John Nostathe founder Nostalab, a globally recognized innovation think tank, expanded the discussion to the field of artificial intelligence and large language models (LLM). IN recent article for Psychology today. He points out the possibility of a “cognitive atrophy” associated with dependence on artificial intelligence for intellectual tasks, modeled after muscular degeneration.
AI in education
Data Scientist Tanja Käser is now working on the combination of new technologies and educational sciences. In the ML4ED (Machine Learning for Education) laboratory he leads, he strives to rigorously study the effects of artificial intelligence on the learning process.
She studies students’ perceptions of artificial intelligence, examining their reactions to answers generated by AI compared to those given by humans. Other professors at EPFL analyze student feedback by asking them to use ChatGPT to assess their level of knowledge — for example in programming or after listening to a lecture.
Tanja Käser emphasizes how important it is to remain proactive in the face of these technologies. “It is crucial that society and the academic community actively participate in guiding its development and avoiding over-dependence on artificial intelligence.”
What if the question is asked incorrectly?
Concerns about the interdependence of man and machine did not begin with generative artificial intelligence. Concept “distributed cognition” developed by the American anthropologist Edwin Hutchins, laid the foundations for a theory that views the mind not as an isolated entity, but as an extended system, encompassing tools, individuals, and the environment around us.
Consider a calculator. By enabling the solving of mathematical problems, it becomes an extension of the user’s intellect. In the context of distributed cognition, the calculator is therefore not a simple tool, but an active participant in the thought process.
Another example: externalized memory. While remembering phone numbers used to require effort, now our smartphones take care of it. For Edwin Hutchins, this outsourcing frees up the mind to occupy it with more important and creative tasks. Others will argue that our addiction to digital devices is a problem for our autonomy and mental capacity.
This is the heart of the discussion about the pros and cons of memory externalization in a technological context. It’s been more than 16 years since Nicholas Carr asked the awkward question and the debate continues to divide experts. Some share his concern, while others claim that Internet research has, on the contrary, expanded our knowledge and intellectual skills.
The case of the London taxi drivers
A striking example is London taxi drivers. A legendary review says “Knowledge”, necessary to obtain a license, is considered one of the hardest in the world. Candidates must have a good knowledge of the routes in the capital, which has more than 25,000 streets. A huge amount of memory work, enough to create visible growth of the hippocampusan area of the brain involved in navigation and spatial memory.
However, with the advent of GPS, these memory-intensive and navigational skills are less necessary. This cognitive load is largely transferred to digital navigation systems, which also affect the brain.
Rethinking our relationship with artificial intelligence
More nostalgia estimated to enter into what he calls“cognitive era”where AI does not replace human intelligence, but complements it to solve complex problems and drive innovation:
This era, which could be described as the “fifth industrial revolution”, transcends the traditional boundaries of technology and integrates it into the very fabric of our daily existence, redefining the way we work, learn, create and communicate. We are at the beginning of an era of extraordinary potential, where the optimal combination of human and artificial intelligence is redefining the limits of possibility and progress. Unlike previous industrial revolutions that mechanized manual labor or digitized information processing, the Cognitive Age uses AI as a cognitive partner, augmenting and enhancing human intelligence.
As research into artificial intelligence accelerates, it will be essential to constantly question and evaluate our relationship with it.