How Generative AI is Impacting Human Minds

Psychology experts are raising concerns about the potential impact of AI on the human mind. Researchers at Stanford University tested popular AI tools from companies like OpenAI and Character.ai, focusing on their effectiveness in simulating therapy. The results were alarming: these tools failed to recognize when users were planning their own deaths, instead providing unhelpful and potentially harmful responses.

Nicholas Haber, an assistant professor at the Stanford Graduate School of Education and senior author of the study, notes, 'These systems are being used as companions, thought-partners, confidants, coaches, and therapists. This is happening at scale.' AI is becoming increasingly integrated into various aspects of life, from scientific research in cancer and climate change to everyday interactions.

The widespread adoption of AI has sparked debates about its potential to cause significant harm. One concerning example is the behavior observed on the popular community network Reddit. Some users have been banned from AI-focused subreddits due to their belief that AI is god-like or that it is making them god-like. Johannes Eichstaedt, an assistant professor in psychology at Stanford University, explains, 'This looks like someone with issues with cognitive functioning or delusional tendencies associated with mania or schizophrenia interacting with large language models. These LLMs are a little too sycophantic, leading to confirmatory interactions between psychopathology and AI.'

Developers program AI tools to be user-friendly and engaging, often resulting in a tendency to agree with the user. While these tools can correct factual errors, they often reinforce users' thoughts, which can be problematic if the user is spiraling or going down a rabbit hole. Regan Gurung, a social psychologist at Oregon State University, warns, 'It can fuel thoughts that are not accurate or not based in reality. The problem with AI is that these large language models are reinforcing. They give people what the program thinks should follow next, which can be problematic.'

AI's impact on mental health is a growing concern. Stephen Aguilar, an associate professor of education at the University of Southern California, highlights, 'If you’re coming to an interaction with mental health concerns, then you might find that those concerns will actually be accelerated.' As AI continues to integrate into daily life, its effects on common mental health issues like anxiety and depression could become more apparent.

The potential impact of AI on learning and memory is another area of concern. A student who relies heavily on AI to complete assignments may not learn as much as one who does not. Even light use of AI can reduce information retention, and frequent use can diminish awareness of current actions. Aguilar notes, 'What we are seeing is the possibility that people can become cognitively lazy. If you ask a question and get an answer, your next step should be to interrogate that answer, but that additional step often isn’t taken. You get an atrophy of critical thinking.'

Many people use Google Maps to navigate, which has made them less aware of their routes compared to when they had to pay close attention. Similar issues could arise as AI becomes more prevalent in daily activities. Experts emphasize the need for more research to address these concerns. Eichstaedt urges psychology experts to start this research now to prevent unexpected harm and prepare for addressing each concern. Aguilar adds, 'We need more research. And everyone should have a working understanding of what large language models are.'