JENSEN LA VENDE

Se­nior Re­porter

jensen.lavende@guardian.co.tt

With the grow­ing use of ar­ti­fi­cial in­tel­li­gence (AI), the Ju­di­cia­ry has is­sued a prac­tice di­rec­tion on the re­spon­si­ble use of AI in court pro­ceed­ings.

In a two-page me­dia re­lease yes­ter­day, the Ju­di­cia­ry gave di­rec­tion on gen­er­a­tive AI us­age and es­tab­lish­ing a frame­work for the re­spon­si­ble and eth­i­cal use of AI in court pro­ceed­ings, which was is­sued by Chief Jus­tice Ron­nie Boodooos­ingh and took ef­fect on Wednes­day.

The guide stat­ed that court pro­ceed­ings must re­main hu­man-con­trolled and that AI can­not re­place in­de­pen­dent le­gal or ju­di­cial judg­ment.

It al­so lim­it­ed the use of AI for draft­ing and ad­min­is­tra­tive pur­pos­es, which will be sub­ject to ver­i­fi­ca­tion and le­gal ci­ta­tions.

The guide­lines come af­ter High Court Judge West­min James was forced to send two at­tor­neys to the Dis­ci­pli­nary Com­mit­tee of the Law As­so­ci­a­tion (LATT) in May last year for sub­mit­ting fic­ti­tious le­gal au­thor­i­ties in a law­suit over the dis­missal of a lab as­sis­tant in 2023.

“Ir­re­spon­si­ble use of in­ter­net sources or gen­er­a­tive AI tools un­der­mines not on­ly in­di­vid­ual cas­es but al­so the cred­i­bil­i­ty of the le­gal sys­tem as a whole. If such con­duct is not con­demned and ap­pro­pri­ate­ly ad­dressed, it could lead to a dan­ger­ous ero­sion of the rule of law,” James said.

The guide­lines al­so pro­hib­it the use of AI to gen­er­ate ev­i­dence, in­clud­ing af­fi­davits and wit­ness state­ments, ex­cept for lim­it­ed for­mat­ting and lan­guage as­sis­tance and re­quire dis­clo­sure when­ev­er AI has been used in prepar­ing court doc­u­ments, in­clud­ing cer­ti­fi­ca­tion of the AI tool used and con­fir­ma­tion that all le­gal au­thor­i­ties have been in­de­pen­dent­ly ver­i­fied.

Judges, Mas­ters, Mag­is­trates and oth­er ju­di­cial of­fi­cers are re­strict­ed from us­ing AI un­less con­duct­ing ad­min­is­tra­tive du­ties and re­search func­tions.

Dur­ing a judg­ment ear­li­er this week, High Court Judge Frank Seep­er­sad stressed the im­por­tant role AI can play in the jus­tice sys­tem and, like Boodoos­ingh, warned that the hu­man el­e­ment must not be ig­nored.

“Mod­ern tech­nolo­gies, in­clud­ing ar­ti­fi­cial in­tel­li­gence, may as­sist judges in le­gal re­search, the iden­ti­fi­ca­tion of au­thor­i­ties, and the or­gan­i­sa­tion of ev­i­dence. Such as­sis­tance en­hances ju­di­cial ef­fi­cien­cy. It does not, how­ev­er, dis­place the con­sti­tu­tion­al func­tion of the judge, whose oblig­a­tion it is to eval­u­ate the ev­i­dence,” Seep­er­sad said in his judg­ment.

In ap­prov­ing the new guide­lines, the Chief Jus­tice said: “We should not wish AI to be de­cid­ing cas­es for us. It can pro­vide use­ful da­ta for us to make hu­man de­ci­sions. The ques­tion is not whether AI should be used, but how can we re­spon­si­bly use AI as a tool to as­sist us in mak­ing the hu­man judg­ments that we need to make?”

Pres­i­dent of the Crim­i­nal Bar As­so­ci­a­tion Se­nior Coun­sel Is­rael Khan wel­comes the guide­lines, say­ing AI can be used to make up judg­ments.

“I will sup­port the Chief Jus­tice; the Crim­i­nal Bar As­so­ci­a­tion will sup­port the Chief Jus­tice in this be­cause if you’re talk­ing about self-de­fence or provo­ca­tion or di­min­ished re­spon­si­bil­i­ty or in­san­i­ty, what­ev­er con­cept you sub­mit­ted, you have to first of all un­der­stand the law in re­la­tion to that and the fac­tu­al ma­trix and make sub­mis­sions on it. So he’s very cor­rect about it, but there was no need for him to, well, maybe there was a need for him to do it in or­der to stem the tide,” Khan said.

Guardian Me­dia con­tact­ed LATT pres­i­dent Lynette See­baran-Suite on the mat­ter; how­ev­er, up to press time there was no re­sponse.