The impact of AI on our way of thinking
The widespread use of generative artificial intelligence tools is changing the way we approach problems and think. Developing a strong awareness of how artificial intelligence algorithms work will become the new priority to avoid falling into the trap of cognitive bias.
In the beginning were "System 1" and "System 2."
Psychologist and Nobel laureate Daniel Kahneman theorizes the Dual Process Theory according to which the human mind has two distinct systems of thought: System 1, intuitive, automatic, fast is responsible for impulsive and instinctive decisions; System 2, rational, deliberative, slow, intervenes to evaluate and correct impulsive decisions.
This model of thinking dual that has dominated cognitive psychology today is beginning to fade because of the new relationship established toward artificial intelligence thanks to the many Large Language Models (LLMs) widely available for public use, such as OpenAI's ChatGPT.
As the capabilities of artificial intelligence increasingly overlap with those of humans, the integration of human labor with artificial intelligence poses fundamental new challenges and opportunities, particularly in knowledge-intensive domains.
The benefits of the introduction of artificial intelligence, especially generative intelligence, are many, especially in terms of efficiency in professionals (DeStefano et al., 2022; Balakrishnan et al., 2022; Valentine and Hinds, 2023), with reductions in repetitive work and the average time to finalize certain tasks. Some studies show efficiency gains in writing (Noy and Zhang, 2023) and scheduling activities (Peng et al., 2023), as well as in ideation and creative work (Boussioux et al., 2023; Girotra et al., 2023).
But massive reliance on AI models, even in everyday life, is not without its critical issues, including "dependencies" and bias, if lacking is a digital literacy focused on understanding and functioning of artificial intelligence models.
A recent article by Massimo Chiriatti and Giuseppe Riva introduced the concept of System 0 in relation to Kahneman's theory of dual thinking. Today, in fact, the constant interaction with artificial intelligence is creating an artificial cognitive system that stands in support of our natural mental processes. That is, it is changing the way we think.
The System 0 is a "layer of distributed intelligence that constantly interacts with our thinking, modifying and enhancing it." In fact, every time we ask ChatGPT to solve a problem, or Spotify about new music to listen to, we tap into System 0, which continuously learns from our interactions and adapts to our preferences, analyzing the data generated by our actions, processing it and producing outputs. Since the "content" of the system is limited by the algorithm's interaction with the human agent, each interaction will incorporate data that will influence future System 0 responses and, consequently, interactions with Systems 1 and 2.
To take full advantage of the potential of technology and not get trapped in its limitations, it is essential to know how the models through which these models operate work. Indeed, System 0 can compensate for some of our natural cognitive limitations by identifying biases in our reasoning, or by suggesting alternative perspectives or reminding us of details we might forget. At the same time, one of the risks is to end up in a "bubble" in which we are exposed to a limited view in which we are exposed to information that tends to confirm already existing preferences.
The role of AI literacy
Critically evaluating and consciously using AI-based technologies will increasingly be a core competency.AI Literacy includes the ability to:
- Know the working principles of algorithms.
- Evaluate the reliability and limitations of the results produced.
- Recognize the ethical and social implications of using AI.
Promoting AI literacy programs, especially for corporations is critical not only for conscious, ethical, and inclusive use of artificial intelligence, but to be able to harness the benefits of AI by empowering their employees to improve performance and their ability to access the potential of System 0, without unknowingly reproducing bias or algorithmic manipulation.
Even Design Thinking, as a methodology that is explicitly human-centered, is not excluded from these changes and can be enhanced by AI. Collaboration between teams of people and AI makes the process faster, more exploratory, and more accessible, but without a solid understanding of AI models, there is a risk of accepting solutions that are not very innovative, reinforcing existing biases in the data or prompts sent, and losing control over the most important decisions. This is whyAI literacy is becoming a fundamental skill for those involved in innovation. It is not enough to know how to use the tools: we need to understand how they work in order to enhance and not simply repeat our initial knowledge.