Artificial Intelligence: Studies Warn AI Could Weaken Memory and Critical Thinking

Artificial Intelligence is transforming the way people work, study and solve everyday problems, but researchers increasingly warn that excessive dependence on generative AI could weaken essential cognitive abilities, including memory, critical thinking and decision-making.

AI-powered chatbots such as ChatGPT, Claude and Gemini can write emails, generate computer code, translate languages, organise travel plans and answer complex questions within seconds. Their growing popularity across schools, universities, workplaces and daily life has sparked concerns about how the technology may affect human thinking over time.

Recent scientific studies suggest that regularly relying on AI for cognitive tasks may reduce people’s ability to learn independently. Researchers identified memory, analytical thinking and decision-making as the mental functions most vulnerable to overdependence on AI systems.

SMA Screening: England to Test All Newborn Babies for Spinal Muscular Atrophy

One American-British study involving 1,222 participants, which is currently undergoing peer review, found that AI tools improved users’ performance in arithmetic and reading comprehension exercises in the short term. However, participants showed weaker long-term learning outcomes and became less willing to solve problems independently when AI tools were unavailable.

The researchers warned that reduced persistence could significantly affect long-term skill development because perseverance remains one of the strongest predictors of successful learning.

Grace Liu, a doctoral researcher at Carnegie Mellon University and the study’s lead author, said AI removes valuable learning opportunities by completing intellectual tasks that traditionally help people build reasoning and problem-solving skills.

Unlike earlier technologies such as calculators, which assisted with calculations while leaving the reasoning process to users, generative AI can perform a wide range of cognitive tasks across multiple disciplines.

Researchers have also highlighted the growing phenomenon of “cognitive offloading,” where people increasingly rely on technology instead of their own mental abilities. Some experts describe the trend as “cognitive surrender,” reflecting a deeper dependence on AI-generated answers.

A 2025 study by the Massachusetts Institute of Technology (MIT) found that students who used generative AI to write essays demonstrated lower levels of critical thinking than those who completed assignments independently.

Johann Chevalere, a researcher in social and cognitive psychology at France’s National Centre for Scientific Research (CNRS), said people naturally seek the easiest way to complete mentally demanding tasks. He warned that AI could reinforce this behaviour by encouraging users to bypass deeper analysis and reflection.

Chevalere explained that the brain strengthens frequently used neural connections while allowing unused pathways to weaken over time. As a result, regularly outsourcing thinking to AI could gradually reduce certain cognitive skills.

In response to these concerns, leading technology companies have started introducing features that encourage users to think independently. OpenAI has launched Study Mode in ChatGPT, while Google offers Guided Learning through Gemini. Instead of immediately providing answers, these tools ask questions and offer hints to help users develop their own solutions.

Microsoft said it has also introduced safeguards in Copilot, including reminders for users to verify AI-generated information and warnings that encourage critical evaluation of responses.

Despite growing concerns, researchers emphasise that evidence remains limited. They say scientists still need large-scale, long-term studies to determine how generative AI affects the human brain over time.

Until more evidence becomes available, experts recommend using AI as a tool to support learning rather than replace independent thinking and problem-solving.