AI tools promise convenience and speed, but if we allow them to do the thinking for us, our own mental muscles could become weaker. Our ability to focus, remember, reason, and imagine depends on practice. If AI takes over too much of that practice, we may pay a hidden cost.
Why memorize processes, facts, or names if a chatbot can recall them instantly? Researchers have shown with the so-called “Google effect” that people remember less when they expect information to be easily retrievable. Navigation apps are a clear example: following digital directions often reduces our sense of direction. Similarly, relying on AI to calculate, outline, or recall quotes means our brains won’t practice retrieval, a key driver of learning.
A recent study from MIT showed another concerning effect of AI tools: they decrease activity in our brains while we work. The study compared brain activity in people who wrote an essay with and without AI and found that the AI users had less brain wave activity while they wrote. Furthermore, the AI users remembered less of their writing afterward. The effects of using AI also got worse over time, leading to long term declines in people’s cognitive activity during writing tasks.
AI might also harm our creativity. Boredom and friction often lead to inventive and creative solutions, but AI could take both of these elements out of our work. One study found that people who were forced to engage in a boring task, like copying numbers from a phone book, came up with more creative solutions for new problems afterward. However, AI tools instantly generate solutions, and users don’t have to spend the time thinking through new ideas. When we only rely on AI, the ideas we generate may have limited creativity. AI tools average across huge datasets, producing “safe” answers that sit near the center of the curve. When we lean too heavily on AI, we might write, design, and brainstorm within its narrow lanes rather than coming up with innovative and human solutions.
Another issue with AI tools is their so-called ‘sycophancy.’ AI systems are often optimized to provide answers that users find appealing right away, often praising the user no matter what. As humans, we need constructive feedback instead of blind praise or affirmation in order to learn and grow. You wouldn’t want a friend or parent who lies to you and only tells you what you want to hear. With AI, we have a tool that is optimized to do just that. After all, these are consumer products. They are made to make us happy! In some extreme cases, this sycophancy problem may have even worsened mentally ill users’ delusions and mania.
The long term side effects of relying on AI won’t be fully recognized for a while. However, we can look to social media as an example to predict what might happen. Social media algorithms are optimized to satisfy users in the short run, but in the long run are negatively linked with mental health. AI systems are designed the same way, so it won’t be surprising if we also find long term drawbacks to using them.
While optimization, productivity, and mental off-loading for monotonous tasks spark our natural interest for improvement, we also need to consider the hidden costs. What are we giving up by having AI generate content than creating our own? What cognitive skills are we sacrificing for expedience? With emergent technology, we can welcome a world where AI tools and productivity-enhancers help us without limiting our creativity and cognition or spiraling us into sycophantic mania. This is an opportunity to decide the future we want.