How much time did you save with AI today?
Gallup’s latest survey
on AI adoption shows that 40% of polled workers used AI “a few times or more” in the past year, 19% used it a few times per week or more, and 8% said they used it daily. All of these percentages were nearly doubled from the previous year. While the numbers still feel low to me, there is no doubting the rapidly growing adoption of AI in the workforce.
At an enterprise level, there are still many barriers preventing AI from being used by all. The largest concerns center around the areas of trustability of model output, cleanliness of the training data, fear of job loss, and a skills gap. Now we can add another potential issue to that list: erosion of critical thinking skills.
MIT’s Media Lab wanted to explore the effect of Generative AI on our brains and the ability to think over time. Understanding the AI-brain connection is becoming more important, especially as we increasingly use AI for more tasks at work and students with still-developing brains are exposed to AI at younger ages.
MIT studied a group of 54 volunteers aged 18-39 and had them write several essays over a four-week period, similar to those asked for on an SAT test. The volunteers were divided into three groups: one that could use ChatGPT, one that could use the Google search engine, and one that could only use their brains to complete the work. Their essays were evaluated by several humans and an AI engine.
During the writing exercises, the research team took EEG measurements to assess cognitive load from all three groups and saw that the ChatGPT group exhibited the lowest brain engagement. As the experiment progressed, ChatGPT users became less engaged, lazier, and by the end of the experiment, just cut and pasted the work from the AI engine into their essays.
The EEG showed that neural connections in the group using Large Language Models (LLMs) decreased by 47% over the course of the experiment. In addition, 83% of those who used LLMs to do the work could not quote from the essay they submitted mere minutes before. In other words, if their human brains had not created the work, it was forgotten almost immediately after they clicked on the Save button.
Admittedly, this was a very small group of participants in the study, and it has not yet been peer-reviewed, but it raises the question of cognitive debt as a result of becoming reliant on technology to do work for us. This study shows that the brain is like any other muscle in our body and can atrophy quickly if not exercised. Short-term gains in speed and productivity with AI could ultimately come at the cost of long-term brain development.
This is a fascinating topic and one that can’t fully be explored in a 600-word blog. However, as AI begins to permeate all business processes, this early study shines a light on our reliance on LLMs and should make companies rethink automation roadmaps and the evolution of the workforce in ways to work with AI and not ceding all control of work and thinking to AI agents.
Do you want to explore AI for your business? Please Contact Us to request a meeting. Don’t have an information architecture for AI you trust yet? LRS can also help you collect, organize, and analyze your data so that it is business-ready.