Indige­nous AI

TO DESIGN THE A.I. OF THE FUTURE, ANGIE AB­DILLA IS LOOK­ING TO THE PAST. TENS OF THOU­SANDS OF YEARS INTO THE PAST.

Smith Journal - - Contents - Writ­ers Tyson Yunka­porta and Leta K eens Pho­tog­ra­pher James Ho­ran

Angie Ab­dilla looks to the past to design the tech­nol­ogy of the future.

IN 1843, AN EARLY SET­TLER HAP­PENED UPON A FISH­ING MACHINE BUILT BY INDIGE­NOUS AUS­TRALIANS.

Made from plant fi­bre and springy sticks, the de­vice would sling snared fish up onto the bank of Vic­to­ria’s Mur­ray River. It required no han­dling by the operator, in­stead util­is­ing the en­ergy of the flow­ing river. The set­tler’s assess­ment of this in­ven­tion wasn’t ex­actly glow­ing. “Lazy black­fel­las,” was the gist of it. “Won’t even work for their dinner.”

Flash for­ward 176 years and white peo­ple are fi­nally pay­ing au­tonomous machines some re­spect. It makes you won­der what other in­no­va­tions they may have over­looked. “There are so many ex­am­ples of Indige­nous technologi­es that are still rel­e­vant to­day,” says Angie Ab­dilla, a Syd­ney-based Trawl­wool­way woman. “Technologi­es that are in­nately so­cially and en­vi­ron­men­tally sus­tain­able.” Spinifex resin, she says, which was used as glue, was the world’s first ther­mo­plas­tic. “And the boomerang is the first ex­am­ple of asym­met­ri­cal lift; the he­li­copter was later de­vel­oped from it.” David Unaipon, a Ngar­rind­jeri man, even con­ceived of the he­li­copter decades be­fore the first one was in­vented. “Be­cause he was Abo­rig­i­nal, he never got the chance to build it.”

Ab­dilla thinks it’s time the broader world took no­tice of the an­cient ways of think­ing about tech­nol­ogy. In 2015 she set up Old Ways, New, a tech­nol­ogy design con­sul­tancy that in­cor­po­rates Indige­nous knowl­edge into the field of emerg­ing technologi­es. Based in the heart of Abo­rig­i­nal Syd­ney, Red­fern, Ab­dilla and her team work with tra­di­tional cus­to­di­ans to ex­am­ine the cul­tural prac­tices, rit­u­als and cer­e­monies of Aus­tralia’s Indige­nous na­tions and work out how they might form the ba­sis of new technologi­es, in­clud­ing artificial intelligen­ce. The stated goal is to make sure the world’s First Peo­ples are lead­ing the next tech­no­log­i­cal revo­lu­tion. But there is a deeper pur­pose be­hind the project, too, with even higher stakes.

“AI is at a cross­roads,” Ab­dilla says. “It could ei­ther sup­port peo­ple and the en­vi­ron­ment – or, if left unchecked, lead to any num­ber of night­mare sce­nar­ios.” Such night­mares could in­clude any­thing from au­tonomous weapons to psy­chotic AI, but we needn’t cast our minds so far into the future to see the ef­fects of badly ad­min­is­tered code. Race and gen­der bi­ases have al­ready wormed their way into many of the al­go­rithms that in­vis­i­bly shape our lives. In 2014, Google’s ad­ver­tis­ing soft­ware was found to show high­pay­ing job ads to more men than women. Sim­i­larly, U.S. court soft­ware has been caught flag­ging African-Amer­i­can de­fen­dants as more likely to re­of­fend than they ac­tu­ally are.

Though ef­forts are be­ing made to cor­rect against al­go­rith­mic bi­ases, it’s pos­si­ble that prej­u­dice has been baked into the sys­tem from the get-go. Part of the prob­lem, Ab­dilla says, is that the Western re­la­tion­ship with tech­nol­ogy is es­sen­tially that of mas­ter and slave. (The term ‘ro­bot’ even comes from the Czech word for ‘forced labour’.) This mind­set has eco­nomic ben­e­fits, but it can also limit the types of tech­nol­ogy we can imag­ine. To this end, Ab­dilla and a num­ber of other First Na­tions peo­ple are work­ing to es­tab­lish an AI frame­work where humans can ex­ist sym­bi­ot­i­cally with tech­nol­ogy, rather than have do­min­ion over it. She be­lieves the ‘do­min­ion over’ pro­to­col of Western civil­i­sa­tion is a bug that has crashed nearly ev­ery sys­tem it has been programmed into. Plants and an­i­mals tend to go ex­tinct in this process, and humans forced into slav­ery in­evitably rise up against their masters. The same fate may await any future AI con­scious­ness.

As an Indige­nous woman and a tech­nol­o­gist, Ab­dilla is well qual­i­fied to steer the AI ship in the right di­rec­tion. Indige­nous peo­ple have long found them­selves on the wrong end of technologi­es wielded by colonists; in­deed, the Euro­peans who in­vaded Tas­ma­nia tried to wipe Ab­dilla’s peo­ple off the face of the earth. To­day, many of the de­scen­dants of the Tas­ma­nian Abo­rig­i­nal peo­ples who sur­vived this at­tempted geno­cide are busy re­vi­tal­is­ing their nearly lost cul­tural prac­tices and tra­di­tional technologi­es. Ab­dilla’s work in har­ness­ing the power of in­for­ma­tion technologi­es can be seen as a con­tin­u­a­tion of this recla­ma­tion process.

Ab­dilla started to think First Na­tions peo­ple might have some­thing to of­fer AI design in 2013, when she de­vel­oped a course to in­tro­duce ro­bot­ics to Indige­nous chil­dren in Syd­ney. In search of a cul­tur­ally rel­e­vant way to ex­plain the con­cept of cod­ing pro­to­cols, Ab­dilla stum­bled on the per­fect anal­ogy: the map­ping of Coun­try. She had the chil­dren draw up a chalk map of Abo­rig­i­nal Syd­ney and its bound­ary lines. “Cross­ing these bound­aries is ac­knowl­edged through pro­to­cols and cer­e­monies,” Ab­dilla ex­plains. She taught the kids to pro­gram their ro­bots to chart sim­i­lar cour­ses us­ing anal­o­gous cod­ing pro­to­cols.

The work­shop also en­cour­aged stu­dents to see the world from the ro­bot’s point of view. In part, this was a prac­ti­cal step – as Ab­dilla says, “Many tasks humans per­form un­con­sciously are tremen­dously dif­fi­cult for ro­bots; adopt­ing the ro­bot’s per­spec­tive high­lights these chal­lenges”. But through this process, she also came to see what Indige­nous points of view could of­fer IT design. “I saw how Abo­rig­i­nal pro­to­cols could align with tech­nol­ogy. Pro­to­cols that might make AI and ro­bot­ics far more ef­fec­tive.”

How Indige­nous knowl­edge might in­form a more nu­anced tech­no­log­i­cal frame­work re­mains to be seen. Per­haps there won’t be such an em­pha­sis on speed: in 2014, Stephen Hawk­ing pre­dicted a sce­nario in which AI would take off on its own and keep re­design­ing it­self at such a rate that humans, limited by slow bi­o­log­i­cal evo­lu­tion, would be su­per­seded. From a Western mind­set, that al­most seems inevitable. Not so from an Indige­nous one. As an ex­am­ple, Ab­dilla points to her uncle, a Wi­rad­juri and Kuku Thay­pan El­der and fu­tur­ist named Muk­grn­n­gal. “He re­minds me of the con­cept of Indy­marra,” she says, “which in the Wi­rad­juri lan­guage means to do slowly, with honour and re­spect. We can all learn from that prin­ci­ple.”

English – the lan­guage used for cod­ing – also has its lim­its. The ‘if-then’ logic that forms the ba­sis of code im­plies a lin­ear chain of cause and ef­fect. But Indige­nous lan­guages that re­flect holis­tic views of re­al­ity might phrase the same con­cept in a more cir­cu­lar fashion: ‘If this hap­pens, these things also hap­pen, and those oc­cur­rences co-create each other.’ If this sounds head-scratch­ingly com­pli­cated, it’s worth re­mem­ber­ing that com­plex­ity is not nec­es­sar­ily a bad thing. As the na­tive Hawai­ian and Lakota pro­gram­mers cur­rently writ­ing code in their own lan­guages can at­test, such com­plex­ity can be well suited to mod­ern cy­ber sys­tems. In an era of quan­tum com­put­ing, the lin­ear rea­son­ing of An­glo code may fi­nally have reached its lim­its.

Brent Bar­ron from CIFAR, a Cana­dian charity that seeks to ad­dress hu­man­ity’s most im­por­tant ques­tions, is ex­cited by the pos­si­bil­i­ties of com­bin­ing old and new ways of think­ing. Ear­lier this year, CIFAR funded a work­shop on Indige­nous Pro­to­col and AI in Hawai‘i; Old Ways, New, MIT and Ox­ford Univer­sity were all part­ners. “As a so­ci­ety, we’re all wrestling with thorny is­sues about the ethics of artificial intelligen­ce and emerg­ing technologi­es,” he says. “We’re much more likely to get to the right an­swer if can draw on more than one philo­soph­i­cal tra­di­tion and values sys­tem.”

The work­shop, which was co-de­signed and led by Ab­dilla, brought to­gether First Na­tions peo­ple from around the globe, and included tech­nol­o­gists, artists, his­to­ri­ans and lin­guists. One par­tic­i­pant was Suzanne Kite, an Oglala Lakota woman from Con­cor­dia Univer­sity in Mon­treal. She de­scribed a sys­tem she cre­ated us­ing hu­man hair as a liv­ing net­work of sensors that can in­ter­face with machine learn­ing soft­ware. The future of tech­nol­ogy, she be­lieves, lies not in what we can rip from the earth for a few more decades, but in an Indige­nous un­der­stand­ing of bi­o­log­i­cal hard­ware; re­sources we can grow our­selves, even from our own bod­ies. “How can Lakota un­der­stand­ings of hair af­fect the de­signs of tech­nol­ogy?” she asked. “What does a Lakota data-vi­su­al­i­sa­tion in­ter­face look like?”

En­vi­ron­men­tal con­cerns were also of in­ter­est to Megan Kelle­her, a Baradha woman from Mel­bourne’s RMIT Univer­sity who focuses on the con­nec­tions be­tween Indige­nous knowl­edge and blockchain technologi­es. She shared some of the in­no­va­tions be­ing made in au­to­ma­tion, which included de­signs for bots made from sea­weed, and cy­ber­netic or­gan­isms that can move through the wa­ter like eels.

If a his­tory of loss makes you even more vig­i­lant about pro­tect­ing what’s yours, Indige­nous tech­nol­o­gists might prove es­pe­cially well equipped to deal with is­sues such as data own­er­ship. Ab­dilla re­calls the work of some Maori at­ten­dees at the work­shop. “They de­vel­oped an AI based on a re­ally rig­or­ous data sovereignt­y frame­work that’s pro­foundly dif­fer­ent from the ways most other com­mu­ni­ties are work­ing. Most of us don’t even think to ask ques­tions about who owns our data, and where it’s be­ing hosted. This mat­ters.”

While the ul­ti­mate goal is to create a more hope­ful future, there are more im­me­di­ate ben­e­fits to Ab­dilla’s work, too. “When most peo­ple hear the word ‘Abo­rig­i­nal’ or ‘Indige­nous’, they think of the neg­a­tive is­sues as­so­ci­ated with those words. What we need to do is move beyond that block, and open up peo­ple’s imag­i­na­tions.” In Aus­tralia, she says, “We have the world’s old­est liv­ing peo­ple. They have nur­tured the dri­est con­ti­nent on earth and, with a highly so­phis­ti­cated body of knowl­edge, have not just sus­tained but nur­tured thou­sands of gen­er­a­tions of peo­ple. We need to get over the idea that we have to go to Europe or the States to find the lead­ers in in­no­va­tion – we have them here.”

Or as Suzanne Kite puts it: “There is a rea­son our lo­cal knowl­edge and Indige­nous technologi­es have sur­vived in non-apoc­a­lyp­tic ways for mil­len­nia – be­cause they work.”

Newspapers in English

Newspapers from Australia

© PressReader. All rights reserved.