When I write a LinkedIn article (including this one), I use AI. I provide the question, the background, and my perspective. I provide the rough outline including sources and then I ask AI to challenge me. I ask it to offer viewpoints I have not considered. AI can access a wider range of information faster than I can but I need to look at the quality of the information.
I take the revised article and edit it. I might even refine it more after more back and forth. I make the final decision about what appears under my name.
I find that valuable. I also find myself wondering what happens when someone stops doing those things.
For my 401k plan sponsors, this is a practical question. AI is here to stay in the workforce. When an employee uses AI to draft a benefits explanation or summarize a report, the employer gets a finished product. Has the employee also gained the understanding needed to answer questions about it? Does the employee have the experience to question the output?
Finishing the Work and Learning the Work
A study of nearly 1,000 high school mathematics students illustrates the question.
Students using a general AI assistant scored 48% higher during assisted practice than students without it. When they subsequently worked without assistance, they scored 17% lower.
Students using a different AI tutor, designed to provide hints instead of simply supplying answers, largely avoided that disadvantage. Help versus hints.
What struck me was that students did not recognize the disadvantage in their own learning. They believed the assistance had helped them without realizing how poorly they would perform independently.
That finding challenges my own assumption. Making the final decision does not automatically mean I understand everything that went into it.
Can I explain the argument? Can I recognize where the evidence falls short? Could I defend the conclusion without reopening the conversation?
Where Does Our Judgment Go?
Research involving 319 working adults found that greater confidence in AI was associated with less reported critical thinking. Participants also described shifting their effort toward checking information, evaluating responses, and deciding how to use them.
That sounds familiar. AI does some of the searching and drafting. I spend more time assessing what it brings back.
Preliminary brain research raises further questions. An MIT-led essay-writing study found lower measured brain connectivity during AI-assisted writing, along with weaker recall of the writing. It does not establish that AI stunts brain development, but it gives us a reason to examine how we engage with these tools.
There is evidence of better learning, too. A randomized college physics study found that a carefully designed AI tutor produced greater learning in less time than the comparison classroom instruction.
The practical choice is how we use the assistance.
What Employers Can Do
I would start with three expectations.
Ask for an explanation. When an employee brings an AI-assisted recommendation, ask why it applies to your organization. A polished paragraph should be the beginning of that discussion.
Use AI to develop understanding. Ask it to explain an unfamiliar concept, question an initial answer, or provide feedback on an employee’s own attempt. The mathematics research suggests that assistance built around working through a problem can preserve learning better than assistance that supplies the solution.
Keep responsibility clear. Someone should review the sources, check the facts, and be able to explain the final answer. In my practice, that responsibility remains with me.
These are practical ways to put the research to work. They also give employers something more useful to evaluate than how quickly a task was completed.
The Parting Glass
I appreciate having access to research I might otherwise miss. I want that research to expand my understanding.
That requires more than reading the finished draft and approving it. I have to consider the evidence, question the reasoning, and understand the choices.
After the article is finished, I should be able to explain what I learned.
The same question belongs in the workplace: when AI helps us produce better work, are we becoming better equipped to do the next task?
#ArtificialIntelligence #CriticalThinking #SmallBusiness #EmployeeDevelopment #Leadership