Researchers at MIT Media Lab released a study on ChatGPT’s possible impact on critical-thinking skills. They assigned 18 people to write three essays using OpenAI’s ChatGPT, 18 to write three essays using Google search and 18 to write three essays relying solely on their brains with no technological help.
The researchers measured brain activity and found that the ChatGPT group had the weakest brain engagement. They also found that the ChatGPT group had a weaker ability to quote from the essay they just “wrote.” By the third essay, the ChatGPT group applied low effort in their writing, mainly copied and pasted and applied minimal editing.
In one final session, researchers asked participants in the ChatGPT group to switch to the brain-only group and the brain-only group to switch to the ChatGPT group. When participants in the group switched to brain-only, they experienced a better sense of ownership of their essays.
A higher memory recall and higher brain engagement. As people of all ages get exposed to AI (artificial intelligence) programs and tools, the impact and consequences, good and bad, are raising questions.
“I think that studies like this are really important because we don’t know exactly what the impact of AI use is going to be for anyone, let alone children,” Dr. Scyatta Wallace, a developmental psychologist, said to The Final Call.
“As someone who has studied brain development and understanding about child development and learning, it only makes sense to try to understand more about how this would impact learning for kids, and especially how they process information. And maybe there’s benefits and maybe there’s consequences.”
The study’s participants were between the ages of 18 and 39 and were recruited from five universities in the Boston, Massachusetts, area. The essay topics were based on SAT tests.
Researchers acknowledged limitations in the study, such as the limited number of participants and the geographical area. To draw more accurate conclusions, larger studies would have to be done with participants from diverse backgrounds, including age and gender and with other large language models (LLMs).
Although the study has not been peer reviewed and has several limitations, it does spark discussion on the impact of artificial intelligence on the development of critical-thinking skills, especially for children.
Dr. Wallace helps Black youth prepare for careers in science, technology, engineering and mathematics (STEM). She explained that children have an infinite possibility of how their brains could grow and described the brain as being like a tree with different branches.
“Those branches are the neurons, and they come together, and when they come together really well, that means that part of that brain is well-developed and that helps them be able to know more, learn more, maybe be better at a particular type of skill. And so that’s why we always say it’s so important to start early with kids, because that helps them build those branches,” she said.
She believes the study is important “because when you use AI, there’s parts of your brain that you’re not using, because AI is doing it for you.” Adults who have already formed “branches,” risk losing the skills they gained by not putting them to use, but heavy AI use can be dangerous for children who have not formed those branches at all.
The study referenced previous research suggesting a strong negative correlation between AI tool usage and critical-thinking skills. Young people tend to have a higher dependence on AI tools, but this also lowers mental processes like thinking, learning, memory and problem-solving.
Now is not the time for Black people to simply respond to studies like this recent one, but rather, to be proactive by creating an intention of how our community could possibly consider utilizing this technology as a tool in some positive ways, Dr. Wallace explained.
She works with teachers and youth leaders to help them create plans for utilizing and managing technological tools. She urged parents to be in the conversation, too, and find out how they want their children to interact with these tools.
“You want to understand what those policies are. You want to understand what they’re doing about it. You want to understand how they’re tracking everything.
If some of these declines are being seen, we need to know this,” she said. “We don’t want to say it’s the worst thing, but we want to be very clear about taking our time and having a strategy on how we want to use it.”
What are some alternative ways to develop a child’s cognitive skills? If technology is being used in the schools, parents of young children can create a learning environment at home, teaching their child how to write by hand and how to write in cursive.
Parents can have conversations to allow their child to think through information and to voice how they came up with their conclusion. Parents of older children can have their child learn about a topic, write it out and write how they thought through the answer.
“Even if it’s a question like how we’re talking about, what are the impacts of AI, you have your child go and figure out,” even if they are utilizing basic versions of AI like Google search, she said.
“They can get some of that information, but then you still have to make them write out okay, why do you feel like AI is going to benefit the brain or not benefit the brain?
And they have to write out and show you their thought process, because that’s a different way of still showing or going through the process that they may lose when they just tell AI to do it.”
For Dr. Wallace, ethical uses of AI in education comes from intention. “You want to have an intentional understanding of how you’re going to use it. You want to understand exactly what some of the benefits and what some of the consequences are.
And like with anything with ethics, it really depends on your worldview and your values, and then you have to make a decision,” she said. “It’s only ethical when you’re actually asking yourself those questions.
When you just do what everybody else is doing, then that’s when it’s unethical, because you’re just largely going into something and you really don’t know what that impact is going to be.”
She acknowledged the presence of racial bias within AI tools like ChatGPT. The Final Call wrote about some of those biases in a previous edition on the positives and negatives of AI (See Vol. 44 No. 16, “Things to know about Artificial Intelligence”).
But AI is not just a robot, Dr. Wallace said. “It’s still people developing it.”