Raphael John-Lall

While Pres­i­dent Chris­tine Kan­ga­loo has made a call for ac­cel­er­at­ed leg­isla­tive ac­tion on Ar­ti­fi­cial In­tel­li­gence (AI), man­ag­ing di­rec­tor of ACE Strate­gic So­lu­tions Ltd, Ri­car­do Fras­er wants the busi­ness com­mu­ni­ty and the pub­lic in gen­er­al to look at the pros and cons of AI’s im­pact.

The Pres­i­dent made her call dur­ing the cer­e­mo­ni­al open­ing of T&T’s 13th Re­pub­li­can Par­lia­ment two weeks ago, in wel­com­ing the new Gov­ern­ment’s de­ci­sion to es­tab­lish a Min­istry of Pub­lic Ad­min­is­tra­tion and Ar­ti­fi­cial In­tel­li­gence. She warned this should be fol­lowed by laws to pre­vent fu­ture harm.In an in­ter­view with the Busi­ness Guardian, Fras­er start­ed off by defin­ing AI and then gave ex­am­ples of how it is trans­form­ing the lo­cal busi­ness com­mu­ni­ty.

He de­fined AI as be­ing an area of tech­nol­o­gy that al­lows ma­chines to per­form rea­son­ing sim­i­lar to that of a hu­man be­ing.

“In a lot of in­stances, the speed at which ma­chines can utilise their rea­son­ing sur­pass­es the lev­el of that of a hu­man be­ing. So, in or­der to achieve AI, which is used for things like pre­dic­tion and ro­bot­ics, ma­chine learn­ing is the sci­ence of learn­ing about be­hav­iour and find­ing pat­terns in or­der to make pre­dic­tions.”

He then spoke about the lat­est and cut­ting edge tech­nolo­gies in AI and how they are af­fect­ing busi­ness op­er­a­tions in T&T.

“AI does present a lot of op­por­tu­ni­ties for the busi­ness sec­tor. We are a small or­gan­i­sa­tion that tra­di­tion­al­ly may not have been able to com­pete with big­ger or­gan­i­sa­tions with larg­er work­forces. But our or­gan­i­sa­tion be­ing small can lever­age AI, ma­chine learn­ing and these tech­nolo­gies to do more. You are see­ing where or­gan­i­sa­tions have been in­creas­ing their pro­duc­tiv­i­ty and speed of de­liv­ery be­cause of AI. So, their busi­ness process­es are now faster, the qual­i­ty is high­er, and they cost less. Re­mem­ber these AI ma­chines do not sleep nor do they eat.”

He gave ex­am­ples of how AI is trans­form­ing T&T’s busi­ness land­scape, cit­ing the fact that Massy Dis­tri­b­u­tion launched AI soft­ware that is able to do price pre­dic­tion. Fras­er said that ca­pac­i­ty will make the re­tail­er’s in­ven­to­ry sys­tems more ef­fi­cient. That pro­vides Massy and its cus­tomers with a win-win sit­u­a­tion be­cause the su­per­mar­ket chain can low­er prices for its cus­tomers as well as op­ti­mise prof­it mar­gins.

“This some­thing is pos­i­tive and it is be­ing done lo­cal­ly,” he said.

Fras­er then gave the ex­am­ple of how AI is trans­form­ing the bank­ing sec­tor, point­ing out that banks are us­ing AI and they are now able to use more au­toma­tion of their ser­vices.

“So, you see banks ad­ver­tis­ing for peo­ple to qual­i­fy for loans in 60 sec­onds. Oth­er banks are say­ing, qual­i­fy for a loan in a few min­utes. I re­al­ly see that a lot of the ser­vices for mem­bers of the pub­lic will be­come eas­i­er to ac­cess. It im­proves the cus­tomer ex­pe­ri­ence,” the tech­nol­o­gy con­sul­tant said.

Leg­is­la­tion need­ed

Fras­er agreed with T&T’s Pres­i­dent that laws and guid­ance for AI will be ben­e­fi­cial.

“As a busi­ness and tech­nol­o­gy con­sul­tant, we see that this is a good thing to avoid abuse of AI by ac­tors who in­tend to mis­lead the pop­u­la­tion. AI needs to be gov­erned so what you want is eth­i­cal use of AI while we in­no­vate. So, while we do not want peo­ple to take ad­van­tage of tech­nol­o­gy to per­form crim­i­nal tech­nol­o­gy or ac­tiv­i­ty which could mar the rep­u­ta­tion of in­no­cent vic­tims and en­ti­ties, we would like AI to be used in a re­spon­si­ble and eth­i­cal man­ner. I see fu­ture laws as a good thing and I await the leg­is­la­tion.”

At the same time, he warned that fu­ture leg­is­la­tion must not be al­lowed to pre­vent new tech­nolo­gies from evolv­ing in the world of busi­ness and tech­nol­o­gy.

“There needs to be bal­ance so as not to stymie or in­hib­it in­no­va­tion, re­search and de­vel­op­ment and in­no­va­tion.”

He gave ex­am­ples of how AI is be­ing used for wrong­do­ing and why leg­is­la­tion is need­ed to pre­vent these types of crim­i­nal acts.

“A sim­ple ap­pli­ca­tion of AI would be a self-dri­ving ve­hi­cle. What it does is it looks for the land­scape around you and the ma­chine learn­ing will be able to make that de­ci­sion to dri­ve.

“How­ev­er, what we have been see­ing of late is abuse of AI to per­form things like Deep Fakes for ex­am­ple. Deep Fakes is some­thing where AI is used to cre­ate im­ages and videos that will repli­cate the ac­tiv­i­ties of hu­man be­ings and it is of­ten used to fool per­sons and that is very dan­ger­ous.

“We al­so see that AI some­times could be bi­ased as if you train the mod­els im­prop­er­ly it could lead to bias and dis­crim­i­na­tion. If you have too much da­ta of one type and not enough da­ta of an­oth­er type of a mi­nor­i­ty pop­u­la­tion, it could lead to dis­crim­i­na­tion of that mi­nor­i­ty pop­u­la­tion.”

He added that AI is al­so used by cy­ber­se­cu­ri­ty at­tack­ers in or­der to in­crease the po­ten­cy of the at­tacks.

“An­oth­er con­cern is the pri­va­cy of da­ta in AI. Da­ta pri­va­cy is a ma­jor is­sue. Look at gen­er­a­tive AI such as Chat­G­PT and so you have a lot of in­tel­lec­tu­al prop­er­ty chal­lenges with the use of AI and who owns that prop­er­ty. There are in­tel­lec­tu­al prop­er­ties that come with abuse es­pe­cial­ly gen­er­a­tive AI. Al­so, you can use these AI mod­els and they can study the pat­terns of the da­ta in or­der that you can op­er­ate in the re­al world with­out hu­man in­ter­ven­tion. In so do­ing, in time this could, if not done prop­er­ly, it could lead to a pri­va­cy vi­o­la­tion.”

Based on these ex­am­ples he gave, he is con­fi­dent that the Gov­ern­ment and oth­er stake­hold­ers will be vi­sion­ary enough to lay AI-re­lat­ed leg­is­la­tion in Par­lia­ment.

Fras­er said he could see AI leg­is­la­tion hap­pen­ing in the medi­um term, be­cause AI has per­vad­ed many peo­ple’s dai­ly lives and its per­va­sive­ness means there is need for gov­er­nance of AI.

“There are two ways to gov­ern AI, you can use poli­cies, and reg­u­la­tions but you can al­so leg­is­late on the use of AI and maybe for some of these is­sues do have le­gal ram­i­fi­ca­tions and this is per­haps the best way to go about it. Some of these in­dus­tries are not on­ly per­ti­nent to in­dus­tries and com­pa­nies but they do have so­ci­etal im­pact. This is why the Eu­ro­pean Union in 2024 en­act­ed cer­tain AI laws and it would not be a sur­prise if T&T sees the need to fol­low this.”

The new Unit­ed Na­tion­al Con­gress (UNC) Gov­ern­ment has es­tab­lished the Min­istry of Pub­lic Ad­min­is­tra­tion and Ar­ti­fi­cial In­tel­li­gence to over­see the de­vel­op­ment of AI and Fras­er gave a few sug­ges­tions on how the Min­istry can de­vel­op poli­cies in this area.

“What we want in T&T is wide adop­tion of AI. The Min­istry charged with the re­spon­si­bil­i­ty and use of AI to im­prove op­er­a­tional ef­fi­cien­cies, to cre­ate pro­duc­tiv­i­ty and re­al­ly en­cour­age the pub­lic sec­tor and oth­er or­gan­i­sa­tions in T&T to com­pete on the glob­al land­scape. Whilst do­ing so, there are ob­vi­ous ben­e­fits for the cit­i­zens of T&T. We will like to see that the Min­istry as­sist the le­gal arm of the coun­try in de­vel­op­ing such laws that will en­sure that AI will be used in an eth­i­cal way. At the same time, you do not want that the laws are so re­stric­tive that no one is able to car­ry out AI with­out in­vest­ing heav­i­ly in­to tools that gov­ern AI.”

He al­so said that if the laws are strin­gent, one would ex­pect that the Min­istry would be able to pro­vide as­sis­tance to those com­pa­nies that would like to in­vest and de­vel­op AI mech­a­nisms for busi­ness en­ti­ties. The con­sul­tant said it is im­por­tant, as well, to ed­u­cate cit­i­zens about the cor­rect use of AI in re­la­tion to the use of their per­son­al in­for­ma­tion.

“What we al­so see are a lot of com­pa­nies us­ing gen­er­a­tive AI and this is to be en­cour­aged and is to be done in a way that does not ex­pose the in­sti­tu­tion­al da­ta of the com­pa­ny and put it at risk. So, these are some of the poli­cies that I would like to see.”