JavaScript is disabled in your web browser or browser is too old to support JavaScript. Today almost all web pages contain JavaScript, a scripting programming language that runs on visitor's web browser. It makes web pages functional for specific purposes and if disabled for some reason, the content or the functionality of the web page can be limited or unavailable.

Saturday, July 5, 2025

MIT study shows ChatGPT abuse impairs brain

by

11 days ago
20250624

The term “Lud­dite” is used to de­scribe per­sons op­posed to new tech­nol­o­gy. In 1800s Eng­land, the first Lud­dites op­posed new au­to­mat­ed ma­chin­ery be­ing in­tro­duced in­to tex­tile mills for fear of job loss­es and re­duced work­er pay. The last few hun­dred years of tech­no­log­i­cal de­vel­op­ment have been pos­i­tive, with eco­nom­ic growth lift­ing bil­lions out of pover­ty world­wide. How­ev­er, the ad­vent of AI pos­es eth­i­cal chal­lenges that may prove present-day Lud­dites part­ly cor­rect in their fears.

As mil­len­ni­als grow­ing up, our par­ents of­ten told us that video games would rot our brains and make us in­to zom­bies. This has been proven false in cer­tain cir­cum­stances, as shown by the study “Us­ing Video Games to Im­prove Ca­pa­bil­i­ties in De­ci­sion Mak­ing and Cog­ni­tive Skill: A Lit­er­a­ture Re­view” by Rey­nal­do et al.

Video games can be used to train per­sons in unique skills such as fly­ing air­craft in Mi­crosoft Flight Sim­u­la­tor. Video games can en­gage the brain in prob­lem-solv­ing ac­tiv­i­ties and pat­tern recog­ni­tion, the brain is not turned off dur­ing a game ses­sion.

How­ev­er, with the ad­vent of AI Large Lan­guage Mod­els, new tech­nol­o­gy has ar­rived that if mis­used will cause cog­ni­tive de­cline. Per­sons who of­fload all of their writ­ing and crit­i­cal think­ing skills to AI mod­els like Chat­G­PT are at risk, since this de­pen­dence ac­tu­al­ly dis­en­gages the use of the brain.

The mis­use of Chat­G­PT and oth­er AI large lan­guage mod­els over long pe­ri­ods of time can cause brain de­cline, ac­cord­ing to a shock­ing new study based on brain da­ta from EEG tests.

In a mam­moth 200-page pa­per en­ti­tled “Your Brain on Chat­G­PT: Ac­cu­mu­la­tion of Cog­ni­tive Debt when Us­ing an AI As­sis­tant for Es­say Writ­ing Task” by Kos­my­na, Na­taliya, et al, re­searchers at MIT pub­lished their da­ta show­ing that over a four-month pe­ri­od, per­sons who re­lied heav­i­ly on Chat­G­PT for writ­ing and rea­son­ing tasks ex­pe­ri­enced cog­ni­tive de­cline mea­sured by weak­er neur­al con­nec­tiv­i­ty.

Per­sons who re­lied heav­i­ly on Chat­G­PT were com­pared against a “brain on­ly” group who were for­bid­den from us­ing AI for writ­ing and rea­son­ing tasks for the four-month pe­ri­od.

“As the ed­u­ca­tion­al im­pact of LLM use on­ly be­gins to set­tle with the gen­er­al pop­u­la­tion, in this study we demon­strate the press­ing mat­ter of a like­ly de­crease in learn­ing skills based on the re­sults of our study. The use of LLM had a mea­sur­able im­pact on par­tic­i­pants, and while the ben­e­fits were ini­tial­ly ap­par­ent, as we demon­strat­ed over the course of 4 months, the LLM group’s par­tic­i­pants per­formed worse than their coun­ter­parts in the Brain-on­ly group at all lev­els: neur­al, lin­guis­tic, scor­ing,” the study said.

The re­searchers used the EEG (elec­troen­cephalo­gram), a test that mea­sures and records the elec­tri­cal ac­tiv­i­ty of the brain, and found the brains of per­sons who over-used Chat­G­PT had weak­er neur­al ac­tiv­i­ty when com­pared to those who did not use it at all.

In their con­clu­sion, the re­searchers warned that overuse of Chat­G­PT re­duced users’ in­cli­na­tion to crit­i­cal­ly eval­u­ate the AI’s out­put. Chat­G­PT’s out­puts are based on cu­rat­ed train­ing da­ta and its pri­or­i­ties are ar­guably bi­ased as a re­sult.

Per­sons who used Chat­G­PT to write es­says felt less of a con­nec­tion to the work and had a much weak­er abil­i­ty to quote the es­says. This can ar­guably mean worse learn­ing out­comes for per­sons who overuse AI to do school­work.

This study should in­form ed­u­ca­tion­al pol­i­cy in T&T, as stu­dents should be taught to use AI re­spon­si­bly and not fall vic­tim to cog­ni­tive de­cline at a young age by re­ly­ing on Chat­G­PT.

Rec­om­men­da­tions for eth­i­cal AI use

De­spite the dan­gers that this MIT study high­light­ed, it is still crit­i­cal for every­one to know how to use AI re­spon­si­bly to keep up in the mod­ern work­place.

As a writer and an at­tor­ney at law, I am keen­ly aware that Chat­G­PT should be used with cau­tion. I usu­al­ly write im­por­tant let­ters my­self and let Chat­G­PT im­prove the tone by mak­ing let­ters more diplo­mat­ic than my per­son­al writ­ing style may pre­fer.

When us­ing Chat­G­PT for re­search, I al­ways ask it to pro­vide me with links to the re­search pa­pers or cas­es it is sug­gest­ing so that I can read those pa­pers my­self. With­out strict in­struc­tions, Large Lan­guage mod­els have a habit of hal­lu­ci­nat­ing books and au­thors that do not ex­ist.

Chat­G­PT is par­tic­u­lar­ly good at cri­tiquing your ex­ist­ing writ­ing and sug­gest­ing im­prove­ments or per­spec­tives you did not con­sid­er at first. This cri­tique can be used to im­prove es­says or work as­sign­ments. How­ev­er, I strong­ly rec­om­mend rewrit­ing Chat­G­PT’s cri­tique in your own words and not copy­ing and past­ing whole­sale for sev­er­al rea­sons.

It is im­por­tant to be crit­i­cal of what­ev­er an AI out­puts, as it may be fac­tu­al­ly in­cor­rect. Fur­ther­more, be­lieve that rewrit­ing what­ev­er good ideas Chat­G­PT may pro­duce in­to your own words would have the same ef­fect as read­ing a top au­thor on a sub­ject and mod­i­fy­ing their ideas to suit your needs. It should im­prove mem­o­ry and spark some de­gree of crit­i­cal think­ing and there­fore avoid the dan­ger­ous out­comes found in the lat­est MIT study on AI caus­ing cog­ni­tive de­cline.


Related articles

Sponsored

Weather

PORT OF SPAIN WEATHER

Sponsored