The Impact of AI on Critical Thinking

With AI now a truly embedded in our daily lives, researchers are beginning to ask, ‘When AI does the all the work, are we truly engaging in critical thought?’ 

This question has profound implications and as recent research from Microsoft points out:

“Used improperly, technologies can and do result in the deterioration of cognitive faculties that ought to be preserved… deprive the user of the routine opportunities to practice their judgement and strengthen their cognitive musculature, leaving them atrophied and unprepared when the exceptions do arise…

While GenAI can improve worker efficiency, it can inhibit critical engagement with work and can potentially lead to long-term overreliance on the tool and diminished skill for independent problem-solving”. 

Psychological Insights into Critical Thinking with AI 

Researchers from Carnegie Mellon University and Microsoft Research conducted an extensive survey of nearly 1,000 real-world examples among knowledge workers who routinely interact with AI tools. Their findings offer an insight into how our cognitive processes are changing.  

The study revealed the existence of a ‘confidence dynamic’. In other words, the more we trust AI’s capabilities, the less we are inclined to scrutinize its outputs. 

They also found that people who are confident in their own skills tend to engage in a more deliberate evaluation of AI-generated content—even if that means expending extra mental effort.  

The Cognitive Trap: Overreliance and Reduced Vigilance 

The research highlights a potential cognitive trap. As AI systems become more sophisticated and we trust them more and more Arguably, our natural tendency to critically evaluate their contributions decreases at exactly the time when critical oversight is most important.  

This can be understood as a by product of ‘cognitive offloading’, whereby a reliance on external aids impact on the activation of our internal, evaluative mechanisms.  

Psychologists theorise that it reduces users’ engagement in deep, reflective thinking, causing our critical thinking skills to atrophy over time. 

Lev Tankelevitch, a senior researcher at Microsoft Research, has commented that ‘low-stakes’ tasks often receive less scrutiny, which may erode our habit of double-checking information which then seeps into our ‘higher-stakes’ decision making. 

Transforming the Nature of Critical Thinking 

The research argues that the integration of AI is reshaping our critical thinking in three significant ways (and they are not all bad): 

1. From Information Gathering to Information Verification: AI's strength in retrieving and organizing data shifts the cognitive burden onto verifying the accuracy of that information.

   
   

   This change requires an active, evaluative mindset to discern reliable insights from potential errors. 

   
   

2. From Problem-Solving to Response Integration: While AI can generate solutions rapidly, the human mind is tasked with integrating these responses into specific contexts - a process that demands higher-order reasoning and adaptation. 

   
   

3. From Task Execution to Task Stewardship: Knowledge workers are evolving from executing tasks to guiding AI in task completion.

   
   

   This new role necessitates a balance between trusting AI outputs and critically overseeing the process to ensure optimal outcomes. 

Evolving Skills for the AI Era 

One school of thought argues that this transformation of critical thinking does not signal its obsolescence; rather, it highlights an evolution in which thoughtful oversight becomes a key determinant of workplace success.  

This potential shift in how we think has significant implications for skill development. It is clear that in-depth knowledge of a topic remains really important, not only to interpret information but also to verify AI outputs.

However, increasingly it seems that this expertise is not sufficient by itself and must be complemented by competencies in directing, evaluating, and integrating AI-generated insights.  

Personally, I agree with the school of thought that ‘we shouldn’t throw the baby out with the bath water’ and that AI is most effective when it serves as a “thought partner” that challenges and refines our reasoning, rather than as a tool on which we rely 100%.