The de­bate be­tween ar­ti­fi­cial in­tel­li­gence and hu­man in­tel­li­gence is not about choos­ing one over the oth­er; it’s about find­ing the bal­ance that max­i­mizes the strengths of both.

That’s ac­cord­ing to Bre­vard Nel­son, co-founder and man­ag­ing di­rec­tor of Caribbean Ideas Synapse, sit­u­at­ed on Wain­wright Street, Port of Spain.

The com­pa­ny helps clients un­der­stand mod­ern tech­nol­o­gy and oth­er as­pects of their busi­ness.

In an in­ter­view with the Sun­day Busi­ness Guardian, Nel­son said the pro­po­nents of AI ar­gue that it rep­re­sents the fu­ture of ef­fi­cien­cy and in­no­va­tion and that AI sys­tems, pow­ered by ma­chine learn­ing and big da­ta, can process vast amounts of da­ta at speeds faster than hu­mans.

Nel­son out­lined the key ar­gu­ments in favour of AI are:

* Ef­fi­cien­cy and speed: AI can au­to­mate repet­i­tive and time-con­sum­ing tasks, free­ing up hu­man re­sources for more strate­gic roles. For ex­am­ple, AI in fi­nance can han­dle high-fre­quen­cy trad­ing far more ef­fi­cient­ly than hu­man traders;

* Da­ta-dri­ven de­ci­sion mak­ing: AI sys­tems can analyse mas­sive datasets to iden­ti­fy pat­terns and trends, pro­vid­ing in­sights that guide more in­formed de­ci­sion mak­ing. This ca­pa­bil­i­ty is par­tic­u­lar­ly valu­able in fields like health­care, where AI can as­sist in di­ag­nos­ing dis­eases by analysing med­ical im­ages; and

* 24/7 avail­abil­i­ty: AI sys­tems do not re­quire rest, en­abling con­tin­u­ous op­er­a­tion. This is crit­i­cal in in­dus­tries such as cus­tomer ser­vice, where AI-pow­ered chat­bots can re­spond im­me­di­ate­ly to cus­tomer in­quiries around the clock.

The case for hu­man in­tel­li­gence

Nel­son not­ed that the ad­vo­cates of hu­man in­tel­li­gence em­pha­sise the qual­i­ties that make peo­ple unique­ly ca­pa­ble of nav­i­gat­ing com­plex, nu­anced sit­u­a­tions. These in­clude cre­ativ­i­ty, emo­tion­al in­tel­li­gence, eth­i­cal rea­son­ing and the abil­i­ty to think ab­stract­ly—all of which are chal­leng­ing for AI to repli­cate...well, for now.

On cre­ativ­i­ty and in­no­va­tion, Nel­son high­light­ed that hu­mans can think cre­ative­ly, gen­er­at­ing new ideas and con­cepts that go be­yond es­tab­lished pat­terns.

Al­so, the tech­nol­o­gy ex­ec­u­tive said hu­mans can un­der­stand and re­spond to emo­tions in a way that AI cur­rent­ly can­not.

“This emo­tion­al in­tel­li­gence is cru­cial in roles that re­quire em­pa­thy, such as coun­selling, ne­go­ti­a­tion and lead­er­ship. It is im­por­tant to note, though, that there have been sig­nif­i­cant strides in bridg­ing this gap with re­cent mod­els recog­nis­ing fa­cial ex­pres­sions and con­ver­sa­tion­al con­texts, but there is still some way to go to have this in­te­grat­ed in­to our dai­ly rou­tines ful­ly,” he ex­plained.

An­oth­er as­pect is that hu­man in­tel­li­gence is guid­ed by moral prin­ci­ples and eth­i­cal con­sid­er­a­tions. While AI can make de­ci­sions based on da­ta, it can­not nav­i­gate the eth­i­cal com­plex­i­ties that of­ten ac­com­pa­ny those de­ci­sions, said Nel­son.

The ques­tion aris­es, does the ad­vance­ment of AI come at the ex­pense of hu­man in­tel­li­gence, or can both co­ex­ist?

He said while some fear that AI could ren­der cer­tain jobs ob­so­lete, lead­ing to a ze­ro-sum sce­nario, many view the fu­ture as en­com­pass­ing a more syn­er­gis­tic re­la­tion­ship be­tween AI and hu­man ca­pa­bil­i­ties.

Look­ing at com­ple­men­tary strengths, Nel­son un­der­scored that AI and hu­man in­tel­li­gence have dif­fer­ent strengths.

“As shared be­fore, both have their strengths but to­geth­er, they can achieve more than ei­ther could alone.”

Delv­ing deep­er, Nel­son in­di­cat­ed that in­stead of re­plac­ing hu­man in­tel­li­gence, AI can aug­ment it.

Giv­ing an ex­am­ple, he said in health­care AI can analyse med­ical da­ta to sug­gest po­ten­tial di­ag­noses, but it is the doc­tors who use their ex­pe­ri­ence and judg­ment to make fi­nal de­ci­sions.

Nel­son point­ed out that the glob­al work­space has been here be­fore.

“At the cusp of every in­dus­tri­al rev­o­lu­tion, where there’s some ma­jor tech­nol­o­gy in­tro­duced, we’ve had this same de­bate. And his­to­ry has shown us that—whether it’s a steam en­gine, the as­sem­bly line, per­son­al com­put­ers, the in­ter­net, you name it—there is this de­bate.”

“Yes, there has been some dis­place­ment of spe­cif­ic kinds of jobs, but more im­por­tant­ly, there has been a cre­ation of new jobs and new in­dus­tries that didn’t ex­ist be­fore. Roles that in­volve man­ag­ing AI sys­tems, in­ter­pret­ing AI-gen­er­at­ed in­sights, and en­sur­ing eth­i­cal AI use are be­com­ing in­creas­ing­ly im­por­tant,” he de­tailed.

A no­table ex­am­ple of this co­ex­is­tence, Nel­son said is in the fi­nan­cial sec­tor where com­pa­nies use AI al­go­rithms for trad­ing and in­vest­ment strate­gies, but hu­man an­a­lysts still play a cru­cial role in in­ter­pret­ing the da­ta and mak­ing strate­gic de­ci­sions.

“This col­lab­o­ra­tion be­tween AI and hu­man in­tel­li­gence has en­hanced the firm’s pro­duc­tiv­i­ty and de­ci­sion-mak­ing ca­pa­bil­i­ties,” he ar­gued.

He said the fu­ture of pro­duc­tiv­i­ty and scal­ing lies not in choos­ing be­tween AI and hu­man in­tel­li­gence, but in the po­ten­tial of aug­ment­ed in­tel­li­gence.

Some strate­gies for co­ex­is­tence

With task di­vi­sion, Nel­son em­pha­sised the fu­ture lies in as­sign­ing AI to tasks that re­quire speed, pre­ci­sion, and da­ta analy­sis, while re­serv­ing hu­man in­tel­li­gence for cre­ativ­i­ty, strat­e­gy, and emo­tion­al­ly com­plex tasks.

“This di­vi­sion al­lows both AI and hu­mans to op­er­ate at their full po­ten­tial. Liken it to what was done when com­put­ers were in­tro­duced in the work­place,” he said.

What hu­man-AI col­lab­o­ra­tion does Nel­son said is as­sist in de­ci­sion-mak­ing process­es.

“For ex­am­ple, AI can pro­vide da­ta-dri­ven rec­om­men­da­tions, while hu­mans ap­ply con­tex­tu­al un­der­stand­ing and eth­i­cal judg­ment to make the fi­nal call.”

The key to con­tin­u­ous learn­ing and adap­ta­tion, the man­ag­ing di­rec­tor stat­ed, is to in­vest in train­ing pro­grammes that help the work­force adapt to work­ing along­side AI.

“As AI evolves, so too must the skills of the hu­man work­force to en­sure that they can ef­fec­tive­ly man­age and com­ple­ment AI sys­tems,” he added.

AI risk

An ar­ti­cle pub­lished on the In­ter­na­tion­al banker.com web­site last month said there are sev­er­al “mi­cro” risks aris­ing from AI use that af­fect in­di­vid­ual fi­nan­cial in­sti­tu­tions. “Ubiq­ui­tous AI use in the fi­nan­cial sec­tor can ex­ac­er­bate threats to con­sumer pri­va­cy and cy­ber­se­cu­ri­ty. More­over, most AI mod­els have an in­her­ent “black box” na­ture, and their pre­dic­tions can­not be eas­i­ly ex­plained. They may al­so prop­a­gate the bi­as­es of the da­ta on which they are trained. Oth­er con­cerns in­clude the emer­gence of da­ta si­los, mod­el hal­lu­ci­na­tions, and al­go­rith­mic co­or­di­na­tion. GenAI mod­els are al­so prone to the prob­lem of “garbage in, garbage out”: the qual­i­ty of the out­puts of these mod­els is on­ly as good as the un­der­ly­ing in­put da­ta,” ac­cord­ing to the ar­ti­cle.

How­ev­er, it stressed that there are al­so “macro” risks that af­fect the sta­bil­i­ty of fi­nan­cial sys­tems.

“As AI use con­tin­ues to gain mo­men­tum, we should re­main at­ten­tive to the sys­temic risks it can cre­ate. Even with its lim­it­ed ca­pa­bil­i­ties, ear­ly AI use caused flash crash­es and fi­nan­cial in­sta­bil­i­ty. No­table ex­am­ples in­clude the 1987 US stock mar­ket flash crash caused in part by the re­liance on rule-based mod­els by in­sur­ance com­pa­nies,” it fur­ther added.