To make sense of the point-counter-point, jar­gon-rid­dled, self-serv­ing debate over con­trol of the devel­op­ment and dan­gers of AI, think For­mu­la One, and an obser­va­tion by an ancient Greek philosopher.

Cre­ators and pro­po­nents of AI are in thrall to the idea of being  in a race. And like all rac­ers, they seem to rel­ish the dan­ger and embrace the fan­ta­sy they won’t be its victim.
For­mu­la One is both unimag­in­ably faster and safer  than when it began.
It reached those pin­na­cles because fierce com­peti­tors co-oper­at­ed to set rules and stan­dards for the com­mon good.
Rules that grew to include driver’s per­son­al safe­ty equip­ment, con­stant improve­ments ncar con­struc­tion, guard rails, crash bar­ri­ers and grav­el run-off areas reduced the dri­ver death rate to only one fatal­i­ty in a For­mu­la One race since 1994.
That’s in spite of both ini­tial fan oppo­si­tion to improved safe­ty (appar­ent­ly for some, the few­er fatal crash­es the less fun in watch­ing) and the aver­age race speed increas­ing from 140 kms per hour in the 1950s to 270 today.
With a glob­al fan base exceed­ing 800-mil­lion, F1 is now the world’s most pop­u­lar sport­ing series.
The les­son ought to have been obvi­ous to the genius­es dri­ving AI to be what they now warn may be too fast to control.
Instead, Dario Amod­ei  CEO of one of the world’s three top AI firms,told CBS Sun­day Morn­ing: “I think for too long, the indus­try lied to peo­ple about the fact that this tech­nol­o­gy had risks.”
And they are still argu­ing not just over what the  rules and safe­guards should be, and who might police them, but in some cas­es, whether there should even be any.
Maybe the AI com­peti­tors should ask their cre­ations to come up with a bet­ter plan than mim­ic­k­ing the modus operan­di  of the  fos­sil fuel indus­try, which glibly glossed over  the effects of car­bon emis­sions they knew were an exis­ten­tial dan­ger long before it became pub­lic knowl­edge, pol­lu­tion from drilling, oil tanker and pipeline spills and their con­sis­tent refusals to clean up the mess­es, or pay com­pen­sa­tion to vic­tims with­out years of litigation.I
In what a stand-up com­ic would sum up with the quip : “I couldn’t make it up if I tried”, the two indus­tries have already man­aged to join them­selves in a sin­gle, fore­bod­ing head­line:
“Data Center’s Spill of 5,000 Gal­lons of Diesel Forces N.J. Riv­er Cleanup”.
The effect on  the “genius” who claims he can pro­tect us all from our worst fears about AI, was predictable:
“There is a SICK con­spir­a­cy going on against AI and Data Cen­ters, and the only one that is hap­py about it is Chi­na. WHOEVER WINS AI, WINS! „,_”
Pres­i­dent Don­ald Trump’s Truth Social ful­mi­na­tion brought to mind a dis­cus­sion I had in South Africa in the ear­ly 1970s with an ardent apartheid sup­port­er, about reports his coun­try was try­ing to build a nuclear weapon.
When I point­ed out that fall­out from a nuclear strike on its neigh­bours would have adverse effects on South Africa, he replied: “It doesn’t mat­ter, because if we drop the bomb, we win.”
As an intel­lec­tu­al argu­ment, that’s on a par with the increas­ing­ly earnest dis­cus­sions about whether or not AI has, or will soon devel­op, “feel­ings”.

                     AI AND ARISTOTLE

When Aris­to­tle made the obser­va­tion that has become  the col­lo­qui­al­ism “The whole is  greater than the sum of its parts”, he had human group­ings and inter­ac­tions in mind.
 Train­ing AI mod­els, how­ev­er, “requires a huge amount of com­pu­ta­tion, which is why many sys­tems rely on spe­cial­ized chips…”
Ergo, AI is, and will remain a machine that does what humans tell it to do, whether by design or accident.
Nonethe­less, there are those who insist that the human-imi­tat­ing way some of the sys­tems have begun to attack and manip­u­late their fel­low what­ev­er-they-ares, means they are becom­ing “one of us”,  only a lot smarter.
The dan­ger is in accept­ing that to the point where “offend­ing AIs sen­si­bil­i­ties” becomes an issue for dis­cus­sion, fol­lowed soon after by the ardent atten­tion of the “every­thin ‑is-offen­sive-to-some­one-and ‑there­fore-must-be-for­bid­den” brigade.
At that point, debate would descend to AI being vul­ner­a­ble, abused. insult­ed, or racial­ly pro­filed, and hence must be catered to rather than, like HAL, the rogue  com­put­er  in “2001: a Space Odyssey”, dis­con­nect­ed if nec­es­sary for the greater good.
An even  more press­ing prob­lem is that the oppo­nents of  any kind of reg­u­la­tion include ven­ture cap­i­tal­ists who see the tech­nol­o­gy as a way to make even more bil­lions than they already have, dan­gers and con­se­quences be damned.|So let us hope that rather than the naivete of my apartheid-era inter­locu­tor, the would-be mas­ters of AI com­pute a proverb from black Africa: “If you want to go fast, go alone; if you want to go far, go togeth­er.”. 

Com­ments are wel­comed. Click CONTACT on the site header.
To receive e‑mail alerts to new posts, Click SIGN-UP on the header.
To SHARE, click the appro­pri­ate icon below.