Is AI making us less intelligent? This is a question that has sparked debate and concern among experts. A recent study from MIT Media Lab suggests that over-reliance on AI solutions may lead to cognitive decline and a decrease in critical thinking abilities. While the study is small and not yet peer-reviewed, it serves as a cautionary tale, even for AI assistants themselves.
The Gazette delved into this topic, interviewing faculty members from various disciplines to explore the impact of AI on critical thinking. We asked: Can AI enhance or hinder our ability to think critically? And what happens when we become too dependent on this technology?
Unraveling the Human Mind's Superiority
Tina Grotzer, a Principal Research Scientist in Education, highlights an important aspect. Many students use AI without understanding its computational workings, leading to an overconfidence in its output. She emphasizes the need to teach students about their own human minds and their incredible capabilities.
Neuroscience research, as Grotzer explains, shows that while human minds use Bayesian processes, they are 'better than Bayesian' in many ways. Our minds can make quick intuitive leaps and detect critical distinctions that AI struggles with. Grotzer's research found that kindergarteners used strategic information more effectively than a purely Bayesian approach.
In her course, 'Becoming an Expert Learner,' Grotzer aims to help students understand the research on human minds and make the most of their unique cognitive abilities. She encourages a thoughtful comparison between human minds and AI, fostering a deeper appreciation for our incredible cognitive powers.
The Assignment vs. The Learning Process
Dan Levy, a Senior Lecturer in Public Policy, emphasizes that AI's impact on learning depends on how it's used. If students use AI to do the work for them, there's little room for learning. Learning occurs when the brain actively engages in making sense of new information, which doesn't happen when we simply ask for answers.
Levy believes that AI can be a helpful tool when it saves time on mundane tasks, allowing students to focus on deeper learning. He acknowledges the dual nature of AI, with both optimistic and worrying aspects, but stresses the importance of finding ways to collaborate with AI to achieve educational goals.
The Owl of Wisdom: A Positive or Negative Force?
Christopher Dede, a Senior Research Fellow, draws an analogy to the Greek goddess Athena, often depicted with an owl on her shoulder, symbolizing wisdom. He asks if AI can be like the owl, guiding us towards wisdom.
Dede warns against letting AI do our thinking for us. While GenAI is excellent at absorbing data and making predictions, it lacks understanding of human context and emotions. The key, he says, is to use AI as a tool to augment our thinking, not replace it. Using AI to do the same old tasks faster can lead to doing the wrong thing more efficiently.
The Limits of AI: Human Experience and Creativity
Fawwaz Habbal, a Senior Lecturer on Applied Physics, teaches a course on AI and Human Cognition, aiming to differentiate between human and machine intelligence. He emphasizes that AI, despite its data processing and statistical prowess, lacks the ability to create truly innovative solutions. Machines calculate, but they don't have human experiences or the capacity for ethical and moral reasoning.
Habbal worries about students becoming overly reliant on AI. He reminds us that leadership development involves adding new value to society, a human enterprise. AI, he argues, cannot perform deep critical thinking or system analysis. Human challenges are complex and require human solutions.
Navigating the Mental Map: AI and Critical Thinking
Karen Thornber, the Harry Tuchman Levin Professor in Literature, highlights how AI forces us to reconsider critical thinking components. While AI can assist with analysis and problem-solving, it struggles with evaluation and reflection, skills that cannot yet be outsourced.
Thornber warns that using AI in certain ways can diminish our skills, from basic memory and knowledge to higher-order critical thinking. The ease of using LLMs may lead us to avoid challenging mental tasks, making it difficult to develop these skills.
The key, she suggests, is to use AI as an assistant for learning and critical thinking, supporting the development of intentional and conscientious AI literacy. Some critical thinking skills will become even more valuable as they remain beyond AI's reach.
The Impact of Cognitive Labor Tools: A Cautionary Tale
Jeff Behrends, a Senior Research Scholar in Philosophy, expresses concern about the effects of LLMs on critical reasoning skills. He points to existing research showing how tools used during cognitive labor can change the way we work. Taking notes by hand, for example, leads to better recall than typing.
Behrends is less worried about AI as an aid in targeted domains, like medical diagnosis, but cautions against the hype surrounding LLMs as general reasoners. He urges caution before embracing technological trends too enthusiastically, reminding us that human challenges require human solutions.
