A recent study conducted by researchers from Microsoft and Carnegie Mellon University examines the impact of generative AI usage in professional settings on critical thinking abilities.
According to the paper, “If misused, technologies can lead to a decline in the cognitive abilities that we should strive to preserve.”
The study notes that when individuals depend heavily on generative AI in their work, their focus shifts toward checking the adequacy of AI-generated responses rather than engaging in higher-level critical thinking activities such as creating, evaluating, and analyzing information. The paper points out that if humans only step in when AI outputs are lacking, they miss out on regular chances to hone their judgment and develop their cognitive skills, ultimately leaving them unprepared for unique challenges.
In essence, excessive reliance on AI for cognitive tasks may hinder our own problem-solving capabilities when the technology fails.
The research involved 319 participants who utilized generative AI at least once a week in their jobs. They were asked to illustrate three instances in which they employed generative AI, categorized into three primary types: creation (such as drafting a standard email); information (like researching a subject or summarizing lengthy articles); and advice (requesting guidance or generating charts from existing data). Participants were then inquired about their use of critical thinking skills for these tasks and whether AI usage influenced their effort in critical thinking. For each task mentioned, respondents also rated their confidence in themselves, the generative AI, and their ability to assess the AI’s outputs.
Approximately 36% of those surveyed acknowledged applying critical thinking skills to counterbalance potential drawbacks of AI use. One participant indicated her use of ChatGPT for crafting a performance review but took the extra step to verify the AI’s input due to concerns of inadvertently submitting something that could lead to her suspension. Another participant mentioned that he modified AI-generated emails for his boss, whose workplace culture highly values hierarchy and age, to avoid any social missteps. In many situations, participants also cross-verified AI outputs with broader web resources like YouTube and Wikipedia, which may undercut the intended efficiency of using AI.
For employees to effectively address the limitations of generative AI, it’s crucial for them to comprehend these limitations; however, not all participants were aware of AI’s constraints.
The paper notes, “Awareness of the potential harms stemming from GenAI responses can stimulate critical thinking, but only if the user is consciously attuned to these harms.”
Interestingly, the study revealed that individuals with confidence in AI tools exerted less effort in critical thinking than those who placed more trust in their own skills.
While the researchers refrain from declaring that generative AI tools diminish intelligence, their findings indicate that an overreliance on these tools may impair our ability to think independently and solve problems.
Compiled by Techarena.au.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
