Though last week’s Caribbean Me­dia Sum­mit, host­ed in Ja­maica by the Me­dia In­sti­tute of the Caribbean (MIC), As­so­ci­a­tion of Caribbean Me­di­a­Work­ers (ACM), and Press As­so­ci­a­tion of Ja­maica (PAJ) fo­cused pri­mar­i­ly on Ar­ti­fi­cial In­tel­li­gence (AI), me­dia and jour­nal­ism, omi­nous pub­lic pol­i­cy mea­sures loomed in the back­ground.

In fact, while me­dia in­dus­try stan­dards and self-reg­u­la­tion were em­pha­sised through­out as sin­gu­lar­ly ef­fec­tive in ad­dress­ing most con­cerns, nu­mer­ous con­ver­sa­tions over the three days of ac­tiv­i­ties drift­ed seam­less­ly in­to what re­gion­al gov­ern­ments may plan to do about the un­fold­ing phe­nom­e­non of “gen­er­a­tive” AI”—the oc­cur­rence of con­tent pro­duced by AI.

One on­line con­trib­u­tor (there were al­most 200 par­tic­i­pants on the Zoom con­nec­tion and an­oth­er 30 to 40 phys­i­cal­ly present) sur­mised that the re­gion’s cus­tom­ary tech­nol­o­gy “lag” was on dis­play, and I thought out loud that per­haps our de­lay in ful­ly en­ter­ing the process can pro­vide some small space to con­sid­er more care­ful­ly how will­ing we re­al­ly are to en­ter­tain reg­u­la­to­ry arms of the State.

It would al­so pro­vide time to con­sid­er hu­man rights, in­clud­ing free­dom of ex­pres­sion, as an in­dis­pens­able el­e­ment of the un­fold­ing sce­nario.

I have long been of the view that al­most every­thing new to en­ter our post-colo­nial space has con­front­ed a pro­hi­bi­tion de­fault. Few bet­ter is­sues than the ma­jor rev­o­lu­tion of Ar­ti­fi­cial In­tel­li­gence to in­stil the kind of fear re­quired to have of­fi­cial­dom ac­ti­vate reg­u­la­to­ry draw­bridges.

It be­came clear as the sum­mit pro­ceed­ed that even among lib­er­al-mind­ed me­dia pro­fes­sion­als, there was a view that cau­tion on the side of of­fi­cial con­trol ought to be a pre­ferred op­tion.

This left pre­cious few of us stand­ing alone. I had mocked the ad­mon­i­to­ry no­tion that “non-hu­man in­ter­ven­tion” in­to pub­lic val­ues and stan­dards pre­sent­ed grave dan­gers and ad­vised that such in­tru­sions were noth­ing new in our con­text and that per­haps or­gan­ised re­li­gion (rel­a­tive­ly free, as it were, from of­fi­cial sanc­tion) could be cit­ed as be­ing among ac­cept­able ex­cep­tions.

Yes, this was high­ly mis­chie­vous on my part, but I thought that the sor­cer­ous im­pli­ca­tions of AI be­ing de­scribed there could have served on­ly to in­voke the kinds of feel­ings that make us want to ar­rest and jail one an­oth­er … or to shut each oth­er up. So, yes, I con­tin­ue to sub­scribe to the view that few­er reg­u­la­tions on pub­lic com­mu­ni­ca­tion pro­duce su­pe­ri­or re­sults over more laws, codes, and puni­tive of­fi­cial mea­sures.

It is al­so true that to­day’s vir­tu­al re­al­i­ties have ren­dered most forms of reg­u­la­to­ry co­er­cion on ex­pres­sion an anachro­nism of the high­est or­der. If any­thing, sanc­tions uni­lat­er­al­ly de­ter­mined by the ma­jor glob­al plat­forms have been far more ef­fec­tive at im­pos­ing regimes of cen­sor­ship.

Last year, for ex­am­ple, my so­cial me­dia pho­to­graph of raw oys­ters was flagged un­der some opaque re­quire­ment to ad­here to “Com­mu­ni­ty Stan­dards.” Last week, the same strange mes­sage no­ti­fy­ing me of a de­ci­sion to block a post ap­peared when I sub­mit­ted an ex­cerpt from this news­pa­per col­umn! And I knew I was not be­ing naughty.

We cus­tom­ar­i­ly protest these mea­sures, cit­ing flawed al­go­rithms or even AI it­self, but the point I am mak­ing is that, to a large ex­tent, big tech com­pa­nies such as Meta (Face­book), Al­pha­bet (Google), Ama­zon, Ap­ple, and Mi­crosoft have out­grown most gov­ern­ments in in­flu­enc­ing on­line be­hav­iours—many times ir­ra­tional­ly and wrong­ly so.

There is thus a vig­or­ous, nu­anced dis­cus­sion on how best to rein in these glob­al su­per­pow­ers in a man­ner that does not com­pro­mise the right to free­dom of ex­pres­sion of their users.

Ad­di­tion­al­ly, and get­ting back to Mark Lyn­der­say’s “time lag” (yes, it was him), the slow pace of dig­i­tal­i­sa­tion by our coun­tries has cre­at­ed a sit­u­a­tion in which the min­ing of on­line con­tent for pur­pos­es of AI con­tent cre­ation gross­ly un­der-rep­re­sents our ac­tu­al re­al­i­ty.

UN­ESCO’s un­fold­ing Caribbean AI Pol­i­cy Roadmap flags some of these is­sues but takes aim more di­rect­ly at the po­ten­tial de­vel­op­men­tal im­pact of in­formed, eth­i­cal us­age. There are sig­nif­i­cant pro­vi­sos, though, that point to as­pi­ra­tions that are cur­rent de­fi­cien­cies.

These in­clude our abil­i­ty to adapt ef­fec­tive­ly to tech­no­log­i­cal ad­vances, ap­pro­pri­ate gov­er­nance mech­a­nisms for de­vel­op­ment and use, and en­sur­ing that the chal­lenges of sus­tain­able de­vel­op­ment, the cli­mate cri­sis, and so­cio-eco­nom­ic in­equities can be ad­dressed.

But all of this would be a tall or­der in the face of de­fi­cien­cies in the ar­eas of train­ing and ed­u­ca­tion and the ur­gent need to pro­mote greater dig­i­tal lit­er­a­cy among our pop­u­la­tions. These short­com­ings can­not be leg­is­lat­ed in­to res­o­lu­tion. My sup­pos­ed­ly an­ar­chic pre­scrip­tion, by­pass­ing the in­stinct to com­mand and con­trol, ap­pears em­i­nent­ly fea­si­ble un­der the cir­cum­stances.