Are artificial intelligence and human intelligence comparable?

Arti­fi­cial intel­li­gence (AI) is dis­rupt­ing the world as we know it. It is per­me­at­ing every part of our lives, with more or less desir­able and ambi­tious goals. Inevitably, AI and human intel­li­gence (HI) are being com­pared. Far from com­ing out of nowhere, this con­fronta­tion can be explained by his­tor­i­cal dynam­ics inscribed deep with­in the AI project.

AI and HI, as fields of study, have co-evolved. There have been two dis­tinct approach­es since the ear­ly days of mod­ern com­put­ing: evo­lu­tion by par­al­lelism or by dis­re­gard. “The founders of AI were divid­ed into two approach­es. On the one hand, those who want­ed to analyse human men­tal process­es and repro­duce them on a com­put­er, in a mir­ror image, so that the two under­tak­ings would feed off each oth­er. On the oth­er, those who saw HI as a lim­i­ta­tion rather than an inspi­ra­tion. This trend was inter­est­ed in prob­lem solv­ing, in oth­er words in the result and not the process”, recalls Daniel Andler.

Our ten­den­cy to com­pare AI and HI in numer­ous pub­li­ca­tions is there­fore not a cur­rent trend, but part of the his­to­ry of AI. What is symp­to­matic of our times is the ten­den­cy to equate the entire dig­i­tal world with AI: “Today, all com­put­ing is described as AI. You have to go back to the foun­da­tions of the dis­ci­pline to under­stand that AI is a spe­cif­ic tool, defined by the cal­cu­la­tion that is being made and the nature of the task it is solv­ing. If the task seems to involve human skills, we will be look­ing at the capac­i­ty for intel­li­gence. That, in essence, is what AI is all about”, explains Maxime Amblard.

The two major trends men­tioned above have giv­en rise to two major cat­e­gories of AI:

• sym­bol­ic AI, based on log­i­cal infer­ence rules, which has lit­tle to do with human cognition
• con­nec­tion­ist AI, based on neur­al net­works, which is inspired by human cognition.

Maxime Amblard takes us back to the con­text of the time: “In the mid­dle of the 20th cen­tu­ry, the com­put­ing capac­i­ty of com­put­ers was tiny com­pared with today. So, we thought that to have intel­li­gent sys­tems, the cal­cu­la­tion would have to con­tain expert infor­ma­tion that we had pre­vi­ous­ly encod­ed in the form of rules and sym­bols. At the same time, oth­er researchers were more inter­est­ed in how exper­tise could be gen­er­at­ed. The ques­tion then became: how can we con­struct a prob­a­bil­i­ty dis­tri­b­u­tion that will pro­vide a good expla­na­tion of how the world works? It’s easy to see why these approach­es explod­ed when the avail­abil­i­ty of data and mem­o­ry and com­put­ing capac­i­ty increased radically”.

To illus­trate the his­tor­i­cal devel­op­ment of these two branch­es, Maxime Amblard uses the metaphor of two skis advanc­ing one after the oth­er: “Before com­put­ing pow­er became avail­able, prob­a­bilis­tic mod­els were ignored in favour of sym­bol­ic mod­els. We are cur­rent­ly expe­ri­enc­ing a peak in con­nec­tion­ist AI thanks to its rev­o­lu­tion­ary results. Nev­er­the­less, the prob­lem of mak­ing the results com­pre­hen­si­ble leaves the way open for hybrid sys­tems (con­nec­tion­ist and sym­bol­ic) to put knowl­edge back into clas­sic prob­a­bilis­tic approaches”.

For her part, Annabelle Blangero points out that today “there is some debate as to whether expert sys­tems real­ly cor­re­spond to AI, giv­en that there is a ten­den­cy to describe sys­tems that nec­es­sar­i­ly involve machine learn­ing as AI”. Nev­er­the­less, Daniel Andler men­tions one of the lead­ing fig­ures in AI, Stu­art Rus­sell, who remains very attached to sym­bol­ic AI. Maxime Amblard also agrees: “Per­haps my vision is too influ­enced by the his­to­ry and epis­te­mol­o­gy of AI, but I think that to describe some­thing as intel­li­gent, it is more impor­tant to ask how what is pro­duced by the com­pu­ta­tion is able to change the world, rather than focus­ing on the nature of the tool used.”

After the his­tor­i­cal and def­i­n­i­tion­al diver­sions, the fol­low­ing ques­tion aris­es: are AI and HI two sides of the same coin? Before we can come up with an answer, we need to look at the method­olog­i­cal frame­work that makes this com­par­i­son pos­si­ble. For Daniel Andler, “func­tion­al­ism is the frame­work par excel­lence with­in which the ques­tion of com­par­i­son aris­es, pro­vid­ed that we call ‘intel­li­gence’ the com­bined result of cog­ni­tive func­tions”. How­ev­er, some­thing is almost cer­tain­ly miss­ing if we are to get as close as pos­si­ble to human intel­li­gence, sit­u­at­ed in time and space. “His­tor­i­cal­ly, it was John Hauge­land who devel­oped this idea of a miss­ing ingre­di­ent in AI. We often think of con­scious­ness, inten­tion­al­i­ty, auton­o­my, emo­tions or even the body”, Daniel Andler explains.

Con­scious­ness and the asso­ci­at­ed men­tal states seem to be miss­ing from AI. For Annabelle Blangero, this miss­ing ingre­di­ent is sim­ply a ques­tion of tech­ni­cal means: “I come from a school of thought in neu­ro­science where we con­sid­er that con­scious­ness emerges from the con­stant eval­u­a­tion of the envi­ron­ment and asso­ci­at­ed sen­so­ry-motor reac­tions. Based on this prin­ci­ple, repro­duc­ing human mul­ti­modal­i­ty in a robot should bring out the same char­ac­ter­is­tics. Today, the archi­tec­ture of con­nec­tion­ist sys­tems repro­duces fair­ly close­ly what hap­pens in the human brain. What’s more, sim­i­lar mea­sures of activ­i­ty are used in bio­log­i­cal and arti­fi­cial neur­al networks”.

Nev­er­the­less, as Daniel Andler points out, “Today, there is no sin­gle the­o­ry to account for con­scious­ness in humans. The ques­tion of its emer­gence is wide open and the sub­ject of much debate in the sci­en­tif­ic-philo­soph­i­cal com­mu­ni­ty”. For Maxime Amblard, the fun­da­men­tal dif­fer­ence lies in the desire to make sense. “Humans con­struct explana­to­ry mod­els for what they per­ceive. We are ver­i­ta­ble mean­ing-mak­ing machines.”

Despite this well-argued devel­op­ment, the ques­tion of bring­ing AI and HI togeth­er remains unan­swered. In fact, the prob­lem is pri­mar­i­ly con­cep­tu­al and con­cerns the way in which we define intelligence.

A clas­sic def­i­n­i­tion would describe intel­li­gence as the set of abil­i­ties that enable us to solve prob­lems. In his recent book, Intel­li­gence arti­fi­cielle, intel­li­gence humaine: la dou­ble énigme, Daniel Andler pro­pos­es an alter­na­tive, ele­gant def­i­n­i­tion: “ani­mals (human or non-human) deploy the abil­i­ty to adapt to sit­u­a­tions. They learn to solve prob­lems that con­cern them, in time and space. They couldn’t care less about solv­ing gen­er­al, decon­tex­tu­alised problems”.

This def­i­n­i­tion, which is open to debate, has the mer­it of plac­ing intel­li­gence in con­text and not mak­ing it an invari­ant con­cept. The math­e­mati­cian and philoso­pher also reminds us of the nature of the con­cept of intel­li­gence. “Intel­li­gence is what we call a dense con­cept: it is both descrip­tive and objec­tive, appre­cia­tive and sub­jec­tive. Although in prac­tice we can quick­ly come to a con­clu­sion about a person’s intel­li­gence in a giv­en sit­u­a­tion, in prin­ci­ple it’s always open to debate”.

In the end, the issue of com­par­i­son seems irrel­e­vant if we are look­ing for a con­crete answer. It is of greater inter­est if we seek to under­stand the intel­lec­tu­al path we have trav­elled, the process. This reflec­tion high­lights some cru­cial ques­tions: what do we want to give AI? To what end? What do we want for the future of our societies?

These are essen­tial ques­tions that revive the eth­i­cal, eco­nom­ic, leg­isla­tive, and social chal­lenges that need to be tak­en up by the play­ers in the world of AI and by gov­ern­ments and cit­i­zens the world over. At the end of the day, there is no point in know­ing whether AI is or will be like us. The only impor­tant ques­tion is what do we want to do with it and why?

Julien Hernandez