Enkulumweni yakhe eHhovisi Lenxusa laseMelika, uHarris uthe umhlaba udinga ukuqala manje ukubhekana “nobuningi obugcwele” bezingozi ze-AI, hhayi nje izinsongo ezikhona ezifana nokuhlasela okukhulu kwe-cyberattacks noma izikhali ze-bioe ezakhiwe yi-AI.
"Kunezinye izinsongo ezifuna ukuthi senze okuthile, izinsongo ezidala umonakalo njengamanje futhi kubantu abaningi nazo zinomuzwa wokuthi zikhona," esho, ecaphuna isakhamuzi esesikhulile siqale uhlelo lwaso lokunakekelwa kwezempilo ngenxa ye-algorithm ye-AI engalungile noma owesifazane osatshiswa umlingani ohlukumezayo ngezithombe zomgunyathi ezijulile.
I-AI Safety Summit iwumsebenzi wothando luka-Sunak, usomabhizinisi wangaphambili othanda ubuchwepheshe ofuna i-UK ibe yisizinda sokusungula amakhompiyutha futhi uye wasungula le ngqungquthela njengesiqalo sengxoxo yomhlaba wonke mayelana nokuthuthukiswa okuphephile kwe-AI.
UHarris kulindeleke ukuthi ahambele ingqungquthela ngoLwesine, ejoyina izikhulu zikahulumeni ezivela emazweni angaphezu kwamashumi amabili nambili okuhlanganisa iCanada, iFrance, iJalimane, i-India, iJapan, iSaudi Arabia - neChina, zimenywe ngemibhikisho yamanye amalungu e-Conservative Party ebusayo iSunak.
Ukuthola izizwe ukuthi zisayine isivumelwano, esibizwa ngokuthi i-Bletchley Declaration, kwaba impumelelo, ngisho noma kukhanya ngemininingwane futhi ayiphakamisi indlela yokulawula ukuthuthukiswa kwe-AI. Amazwe azibophezele ukusebenzela “isivumelwano esabiwe kanye nesibopho” mayelana nobungozi be-AI, futhi abambe uchungechunge lweminye imihlangano. ISouth Korea izobamba ingqungquthela ye-AI encane ye-AI ezinyangeni eziyisithupha, ilandelwe eyokuba seFrance ngonyaka kusukela manje.
IPhini likaNgqongqoshe Wezesayensi Nobuchwepheshe eChina, uWu Zhaohui, uthe ubuchwepheshe be-AI “abunasiqiniseko, abuchazeki futhi abukho obala.”
"Kuletha ubungozi kanye nezinselelo ezimisweni zokuziphatha, ukuphepha, ubumfihlo kanye nokungenzeleli. Ubunkimbinkimbi bakho buyavela," esho, ephawula ukuthi uMongameli waseChina u-Xi Jinping ngenyanga edlule wethule i-Global Initiative yezwe ye-AI Governance.
"Sicela ukubambisana komhlaba wonke ukwabelana ngolwazi nokwenza ubuchwepheshe be-AI butholakale emphakathini ngaphansi kwemibandela yomthombo ovulekile," esho.
I-CEO yakwaTesla u-Elon Musk naye uhlelelwe ukuxoxa nge-AI noSunak engxoxweni ezosakazwa ngoLwesine ebusuku. Usozigidigidi wezobuchwepheshe wayephakathi kwalabo abasayine isitatimende ekuqaleni kwalo nyaka bephakamisa i-alamu mayelana nezingozi i-AI ezibeka esintwini.
UMongameli weKhomishini YaseYurophu u-Ursula von der Leyen, uNobhala-Jikelele weNhlangano Yezizwe u-Antonio Guterres kanye nabaphathi bezinkampani zezobunhloli zokwenziwa zase-US ezifana ne-Anthropic, i-DeepMind yakwaGoogle kanye ne-OpenAI kanye nososayensi bamakhompiyutha abanethonya abafana no-Yoshua Bengio, ongomunye “wonkulunkulu” be-AI, nabo bahambele umhlangano obuseBletchley Park njengendawo yokuzalwa yesimanjemanje ye-World War II. ikhompuyutha.
Abahambele lo mhlangano bathe indlela yomhlangano ovaliwe ibikhuthaza inkulumompikiswano enempilo. Izikhathi zokuxhumana ezingakahleleki ziyasiza ekwakheni ukwethembana, kusho uMustafa Suleyman, oyi-CEO ye-Inflection AI.
Khonamanjalo, ezingxoxweni ezisemthethweni "abantu bakwazile ukwenza izitatimende ezicace kakhulu, futhi yilapho ubona khona ukungaboni ngaso linye okukhulu, kokubili phakathi kwamazwe asenyakatho naseningizimu (kanye) namazwe avuna umthombo ovulekile futhi kancane avuna umthombo ovulekile," u-Suleyman etshela izintatheli.
Amasistimu e-AI yomthombo ovulekile avumela abacwaningi nochwepheshe ukuthi bathole izinkinga ngokushesha futhi babhekane nazo. Kodwa okubi ukuthi uma uhlelo lomthombo ovulekile selukhishiwe, "noma ubani angalusebenzisa futhi aluculele izinjongo ezinonya," kusho uBengio eceleni komhlangano.
"Kunalokhu kungahambisani phakathi komthombo ovulekile nokuphepha. Ngakho-ke sibhekana kanjani nalokho?"
Ohulumeni kuphela, hhayi izinkampani, abangagcina abantu bephephile ezingozini ze-AI, kusho uSunak ngesonto eledlule. Kodwa-ke, uphinde wanxusa ngokumelene nokujaha ukulawula ubuchwepheshe be-AI, ethi kufanele kuqondwe ngokugcwele kuqala.
Ngokuphambene, uHarris ugcizelele isidingo sokubhekana nalokhu lapha namanje, okuhlanganisa “nokulimala komphakathi osekwenzeka vele njengokuchema, ukucwasa kanye nokwanda kolwazi olunganembile.”
Ukhombe umyalo kaMongameli uJoe Biden kuleli sonto, ebeka izivikelo ze-AI, njengobufakazi ukuthi i-US ihamba phambili ngesibonelo ekwakheni imithetho yobuhlakani bokwenziwa esebenzela umphakathi.
UHarris uphinde wakhuthaza amanye amazwe ukuthi abhalisele isithembiso esisekelwe e-US sokunamathela ekusebenziseni “isibopho nokuziphatha” kwe-AI ngezinjongo zempi.
"Mina noMongameli Biden sikholelwa ukuthi bonke abaholi ... banomsebenzi wokuziphatha, wezimiso zokuziphatha kanye nomphakathi wokuqinisekisa ukuthi i-AI iyamukelwa futhi ithuthuke ngendlela evikela umphakathi ekulimaleni okungenzeka futhi iqinisekise ukuthi wonke umuntu uyakwazi ukujabulela izinzuzo zayo," kusho yena.
Isikhathi sokuthumela: Nov-21-2023
