Dr Fazal Ali

To shape the fu­ture, we must fash­ion AI sys­tems that prompt us to think bet­ter and more hu­mane­ly than we nat­u­ral­ly do. AI opens up the pos­si­bil­i­ty of aug­ment­ing hu­man in­tel­li­gence and solv­ing prob­lems with wis­dom. In that fer­ment lies the pos­si­bil­i­ty of an ef­flo­res­cence of ma­chine ideas, new cre­ations, even new species of ideas at the in­ter­sec­tion of AI-en­abled neu­ro­pros­thet­ics.

This will not al­ter the fact that these new ideas, and new species of ideas at the in­ter­sec­tion of the cor­tex and neu­ro­pros­thet­ics, are still hu­man ideas, but they have been boost­ed and en­abled by AI to make us bet­ter hu­mans. Hu­mans nev­er think about the world with an emp­ty mind. Our past shapes what we see first.

There is no im­mac­u­late per­cep­tion. Every­thing is al­ready un­der­stood. There is al­ways some “Vorhaben” or fore-mean­ing, fore-un­der­stand­ing, or pre-project that is brought for­ward to every ex­pe­ri­ence and thought. In the In­tel­li­gent Age, hu­mans prompt AI mod­els to achieve a range of ob­jec­tives along a con­tin­u­um from the de­cent to the ne­far­i­ous.

Re­gard­less of where our prompts fall along this con­tin­u­um, we know that hu­man think­ing is im­per­fect and coloured by fore-un­der­stand­ing, and mem­o­ry is a brico­lage of ed­its and for­get­tings.

We must there­fore ask: What kind of think­ing do AI-en­abled roles ex­pect from hu­man de­lib­er­a­tive ra­tio­nal­i­ty?

Cre­ative screen economies need thinkers who dis­play blind naivety. Work­ers who chal­lenge as­sump­tions. They are the ones who will be cru­cial for or­gan­i­sa­tion­al agili­ty and in­no­va­tion.

AI-flu­en­cy does not erode this ca­pac­i­ty. But AI-il­lit­er­a­cy can wear away chances of find­ing jobs in AI-en­abled roles. The fu­ture jobs mar­ket wants da­ta-flu­ent work­ers. Work­ers who are will­ing to be prompt­ed by AI to think bet­ter and more hu­mane­ly than we nat­u­ral­ly do. Work­ers who can su­per­vise AI.

AI-as­sem­blages are forc­ing firms to fo­cus on two or­gan­i­sa­tion­al risks re­lat­ed to the ero­sion of crit­i­cal think­ing: the for­fei­ture of ex­per­tise and a deficit of epis­temic plu­ral­ism. These risks harm their abil­i­ty to stay ag­ile and in­no­v­a­tive. Com­pa­nies un­der­stand that AI is not neu­tral. They know that AI is hu­man. They know that we need to shape AI rather than be shaped by AI.

Work­ers who trust an­a­lyt­ic tools with­out un­der­stand­ing them could lose their ex­per­tise. AI can com­pro­mise peo­ple’s abil­i­ty to ques­tion the in­for­ma­tion they re­ceive, rais­ing con­cerns about the harms of cog­ni­tive of­fload­ing and cog­ni­tive sur­ren­der. Un­bri­dled use of AI may al­so lead to the de­con­tex­tu­al­i­sa­tion of ex­per­tise with­in or­gan­i­sa­tions. AI can read­i­ly ac­cess facts and da­ta but miss­es nu­ances.

Nu­ances are recog­nised by peo­ple with in­tu­ition, emo­tion, and lived ex­pe­ri­ence of mul­ti­ple do­mains of ex­per­tise. They see things in the da­ta be­cause they have di­rect ex­pe­ri­ence of life and liv­ing and un­der­stand con­text. An al­go­rithm does not ex­pe­ri­ence the life-worlds of hu­mans di­rect­ly; it over­looks nu­ances.

AI nur­tures ho­mo­ge­neous think­ing and re­duces the di­ver­si­ty of ideas, part­ly be­cause it has a bias to­wards av­er­ag­ing. This en­cour­ages hu­man prompters to ig­nore un­fore­seen in­for­ma­tion that can trig­ger cre­ative in­sights. This leads to the de­vel­op­ment of mono­cul­tures that lack the di­verse ex­pe­ri­ences and per­spec­tives es­sen­tial to an or­gan­i­sa­tion’s ca­pac­i­ty to in­no­vate.

This fur­ther con­strains a com­pa­ny’s agili­ty, as work­ers are more like­ly to miss en­vi­ron­men­tal sig­nals that in­vite change. The task ahead is to de­sign AI-en­abled work en­vi­ron­ments that pre­serve hu­man agency and lead­er­ship. Bu­reau­cra­cies can en­cour­age lo­cal eco­nom­ic de­vel­op­ment units (LEDs) and mu­nic­i­pal cor­po­ra­tions to use AI tools dur­ing spe­cif­ic time blocks or pro­vide them with gat­ed in­ter­faces that un­lock as­sis­tance on­ly af­ter the teams frame their projects and up­load them.

Prob­lem fram­ing is not prob­lem find­ing or prob­lem solv­ing. The same prob­lem framed in dif­fer­ent ways pro­duces dif­fer­ent so­lu­tion strate­gies. An­oth­er ap­proach is to host AI-free strat­e­gy ses­sions. Squads can first brain­storm their strate­gic plans based on their own judge­ment and ex­pe­ri­ence ahead of up­load­ing and prompt­ing AI in­ter­faces.

In some set­tings, two squads can run sprints on the same chal­lenge in par­al­lel. One us­es a tra­di­tion­al de­sign process from be­hind a veil of ig­no­rance, while the oth­er re­lies heav­i­ly on AI-sup­port­ed work­flows.

The AI-sup­port­ed team will usu­al­ly ar­rive at op­tions that dra­mat­i­cal­ly ex­pand the ex­plo­ration space with bur­geon­ing di­ver­gence.

The hu­man-led team will find few­er routes, with stronger co­her­ence, sharp­er nar­ra­tives, and more con­tex­tu­al prob­lem frames. Bring­ing the two teams’ out­puts to­geth­er cre­ates an in­cred­i­ble ef­flo­res­cence of ma­chine ideas and hu­man ideas that are still hu­man ideas in the end.

Good­ness, as ra­tio­nal­i­ty, does not at­tribute any spe­cial virtue to the process of de­cid­ing. Hu­man think­ing is im­per­fect. The AI fron­tier opens the pos­si­bil­i­ty to aug­ment each per­son’s Agili­ty Quo­tient to think bet­ter. In that fer­ment lies the pos­si­bil­i­ty of a blos­som­ing of hu­man ideas boost­ed by AI.

The true chal­lenge is there­fore to build AI sys­tems that prompt us to think more benev­o­lent­ly than we nat­u­ral­ly do.

Dr Fazal Ali com­plet­ed his Mas­ter's in Phi­los­o­phy at the Uni­ver­si­ty of the West In­dies. He was a Com­mon­wealth Schol­ar who at­tend­ed the Uni­ver­si­ty of Cam­bridge, Hugh­es Hall; the Provost of the Uni­ver­si­ty of Trinidad and To­ba­go; the act­ing Pres­i­dent of UTT; and the Chair­man of the Teach­ing Ser­vice Com­mis­sion. He is the Pres­i­dent of NI­HERST and an ex­ter­nal ser­vices con­sul­tant with the IDB.