M\u0101ksl\u012bg\u0101 intelekta (MI) tehnolo\u0123iju strauj\u0101 izaugsme p\u0113d\u0113jos gados ir veicin\u0101jusi ar\u012b nopietnas diskusijas par t\u0101s dro\u0161\u012bbu un ietekmi uz sabiedr\u012bbu. \u0160o diskusiju priek\u0161pl\u0101n\u0101 non\u0101kusi Future of Life Institute (FLI) pag\u0101ju\u0161aj\u0101 ned\u0113\u013c\u0101 publiskot\u0101 atkl\u0101t\u0101 v\u0113stule1, kur\u0101 t\u0101s autori br\u012bdina, ka MI izstr\u0101des sacens\u012bba k\u013cuvusi nekontrol\u0113jama, un uz\u0146\u0113mumi steidz rad\u012bt arvien jaud\u012bg\u0101kus “digit\u0101los pr\u0101tus”, kurus vairs nesp\u0113j piln\u012bb\u0101 saprast vai kontrol\u0113t pat to rad\u012bt\u0101ji.<\/p>\n
FLI ir starptautisks bezpe\u013c\u0146as p\u0113tniec\u012bbas instit\u016bts, kura m\u0113r\u0137is ir pan\u0101kt, lai jaun\u0101s tehnolo\u0123ijas kalpotu cilv\u0113ces lab\u0101 un nerad\u012btu glob\u0101lus riskus, \u012bpa\u0161u uzman\u012bbu piev\u0113r\u0161ot progres\u012bva MI potenci\u0101lajiem draudiem. Jau 2023. gada mart\u0101 FLI public\u0113ja pla\u0161u rezonansi izrais\u012bju\u0161u atkl\u0101to v\u0113stuli, kur\u0101 aicin\u0101ja visas pasaules MI laboratorijas uzreiz noteikt vismaz se\u0161us m\u0113ne\u0161us ilgu pauzi t\u0101du MI sist\u0113mu izstr\u0101d\u0113, kas jaudas zi\u0146\u0101 p\u0101rsp\u0113j taj\u0101 br\u012bd\u012b eso\u0161o GPT-4 modelii<\/a>.<\/p>\n Tagad FLI aicina izmantot pauzi, lai izstr\u0101d\u0101tu dro\u0161\u012bbas standartu sist\u0113mu un aizsardz\u012bbas meh\u0101nismus, pret\u0113j\u0101 gad\u012bjum\u0101 b\u016btu j\u0101iejaucas vald\u012bb\u0101m, \u012bstenojot moratoriju uz \u0161\u0101du b\u012bstami sp\u0113c\u012bgu MI izstr\u0101di. V\u0113stul\u0113 uzsv\u0113rts, ka nekontrol\u0113ta MI att\u012bst\u012bba var izrais\u012bt “dzi\u013cu l\u016bzumu cilv\u0113ces v\u0113stur\u0113” un tiek uzskait\u012bti iesp\u0113jamie riski k\u0101 MI \u0123ener\u0113ta dezinform\u0101cija un propaganda, masveida darbvietu automatiz\u0101cija, cilv\u0113ku “aizst\u0101jam\u012bba” un sabiedr\u012bbas m\u0113roga kontroles zaud\u0113\u0161ana p\u0101r notikumiem.<\/p>\n V\u0113stuli \u012bs\u0101 laik\u0101 parakst\u012bja daudzi tehnolo\u0123iju eksperti, p\u0113tnieki un uz\u0146\u0113m\u0113ji, un kop\u0113jais parakstu skaits v\u0113l\u0101k p\u0101rsniedza 50 t\u016bksto\u0161us. Starp redzam\u0101kajiem parakst\u012bt\u0101jiem ir, piem\u0113ram, Nobela pr\u0113mijas laure\u0101ti D\u017eefrijs Hintons un Darons A\u010demoglu, MI p\u0113tnieks Jo\u0161u\u0101 Ben\u017eio, datorzin\u0101tnieks Stjuarts Rasels, uz\u0146\u0113m\u0113ji Ri\u010dards Brensons<\/a> un St\u012bvs Voz\u0146aks, v\u0113sturnieks Juvals Noa Harari, p\u0101vesta MI padomnieks Paolo Benanti un citi. \u0160\u0101da pla\u0161a koal\u012bcija no in\u017eenieriem un akad\u0113mi\u0137iem l\u012bdz uz\u0146\u0113mumu vad\u012bt\u0101jiem un dom\u0101t\u0101jiem skaidri par\u0101da, ka ba\u017eas par MI nekontrol\u0113tas att\u012bst\u012bbas riskiem vairs nav tikai \u0161aura pesimistu loka fant\u0101zijas. T\u0101s ir k\u013cuvu\u0161as par nopietnu dienask\u0101rt\u012bbas jaut\u0101jumu tehnolo\u0123iju nozar\u0113 un sabiedr\u012bb\u0101 kopum\u0101.<\/p>\n Kaut ar\u012b aicin\u0101t\u0101 faktiska pauze MI “sacens\u012bb\u0101” pagaid\u0101m nav ieviesta un visdr\u012bz\u0101k tas ar\u012b nenotiks, FLI pazi\u0146ojums jau ir izrais\u012bjis pla\u0161\u0101kas diskusijas par MI dro\u0161\u012bbu. Tas strauji ieg\u016bst publicit\u0101ti glob\u0101lajos medijos (Latvijas medijos \u0161is notikums gan pasl\u012bd\u0113jis gar\u0101m nepaman\u012bts) un rosina politikas veidot\u0101jus steidzam\u0101k piev\u0113rsties MI p\u0101rvald\u012bbas jaut\u0101jumiem. Vienlaikus MI att\u012bst\u012bba nav apst\u0101jusies, jo uz\u0146\u0113mumi turpina ieguld\u012bt iev\u0113rojamus resursus arvien jaud\u012bg\u0101ku mode\u013cu izstr\u0101d\u0113, kas tikai pastiprina nepiecie\u0161am\u012bbu p\u0113c dro\u0161\u012bbas standartiem. FLI iniciat\u012bva \u0161obr\u012bd pal\u012bdz rad\u012bt “steidzam\u012bbas saj\u016btu” vald\u012bb\u0101s un atspogu\u013co sabiedr\u012bb\u0101 pieaugo\u0161\u0101s ba\u017eas par MI rad\u012btajiem riskiem. Citiem v\u0101rdiem, pazi\u0146ojums jau tagad aktualiz\u0113 politikas dienask\u0101rt\u012bb\u0101 jaut\u0101jumu: k\u0101 nodro\u0161in\u0101t, lai MI att\u012bst\u012bba notiek sabiedr\u012bbas interes\u0113s un dro\u0161\u0101 veid\u0101.<\/p>\n \u0160o piesardz\u012bbas aicin\u0101jumu v\u0113l sare\u017e\u0123\u012bt\u0101ku padara \u0161odienas tirgus dinamika, v\u0113rojot bezprecedenta invest\u012bcijas “superintelekta” att\u012bst\u012bb\u0101. Glob\u0101lo tehnolo\u0123iju uz\u0146\u0113mumu sacens\u012bba par “superintelektu” jau tagad noz\u012bm\u0113 milzu t\u0113ri\u0146us datu centros, \u010dipos un talantos. Neskatoties uz desmitiem miljardu ieguld\u012bjumu vilni, vienlaikus pieaug ba\u017eas par iesp\u0113jamu burbuli, ko atz\u012bst ar\u012b industrijas l\u012bderi (t.sk. OpenAI vad\u012bt\u0101js Sems Altmans)ii<\/a>. Person\u0101la tirg\u016b uz\u0146\u0113mumi c\u012bn\u0101s par zvaig\u017e\u0146u p\u0113tniekiem ar daudzu miljonu atl\u012bdz\u012bb\u0101m, bet atsevi\u0161\u0137u lielo mode\u013cu treni\u0146\u0161 jau p\u0101rsniedz 100 milj. USD, ilustr\u0113jot izmaksu eskal\u0101cijuiii<\/a>. Ar\u012b Meta atkl\u0101ti “lej miljardus” sav\u0101 pl\u0101n\u0101 virz\u012bties uz “person\u012bgo superintelektu”, kas v\u0113l vair\u0101k pastiprina kapacit\u0101\u0161u un kapit\u0101la bru\u0146o\u0161anos starp lielajiem sp\u0113l\u0113t\u0101jiemiv<\/a>. Kopum\u0101 aina ir skaidra – sacens\u012bba starp ASV, \u0136\u012bnu un to vado\u0161ajiem uz\u0146\u0113mumiem piespie\u017e visus palielin\u0101t ieguld\u012bjumus rekordtemp\u0101, kam\u0113r politikas veidot\u0101ji un investori sprie\u017e, vai straujums nav p\u0101rv\u0113rties par riskantu p\u0101rinvest\u0113\u0161anuv<\/a>.<\/p>\n Par FLI pazi\u0146ojumu jau izskan kritika, un koment\u0113t\u0101ji nor\u0101da, ka glob\u0101ls aizliegums ir nerealiz\u0113jams un gr\u016bti izpild\u0101ms apst\u0101k\u013cos, kad valstis un “Big Tech” turpina sacens\u012bbu, turkl\u0101t pazi\u0146ojum\u0101 nav skaidri defin\u0113ts, kas tie\u0161i b\u016btu “superintelekts” un k\u0101 izm\u0113r\u012bt “pla\u0161u zin\u0101tnisku konsensu” vai “sp\u0113c\u012bgu sabiedr\u012bbas atbalstu”. Nozares un da\u013ca politikas veidot\u0101ju iebilst, ka \u0161\u0101ds solis var\u0113tu bremz\u0113t inov\u0101ciju, savuk\u0101rt vair\u0101ki koment\u0113t\u0101ji pied\u0101v\u0101 alternat\u012bvu nevis “aizliegumu”, bet caurskat\u0101m\u012bbu un regul\u0113tu att\u012bst\u012bbu (piem\u0113ram, pras\u012bbas par test\u0113\u0161anu, atbild\u012bbu un risku mazin\u0101\u0161anu).<\/p>\n Ar\u012b iepriek\u0161\u0113j\u0101 FLI v\u0113stule izpeln\u012bj\u0101s kritiku un vair\u0101ki MI jomas eksperti pauda viedokli, ka aicin\u0101jums, iesp\u0113jams, dramatiz\u0113 situ\u0101ciju un nov\u0113r\u0161 uzman\u012bbu no jau eso\u0161aj\u0101m probl\u0113m\u0101m. Piem\u0113ram, MI \u0113tikas p\u0113tniece Timnita Gebru un citi nor\u0101d\u012bja, ka atkl\u0101t\u0101 v\u0113stule piev\u0113r\u0161as t\u0101liem, hipot\u0113tiskiem “distropiskiem sci-fi” scen\u0101rijiem, kam\u0113r re\u0101li \u0161odien past\u0101v riski – aizspriedumi algoritmos, MI rad\u012bta diskrimin\u0101cija, maldino\u0161i automatiz\u0113ti l\u0113mumi, kas prasa t\u016bl\u012bt\u0113jus risin\u0101jumusvi<\/a>. Neskatoties uz to, FLI iniciat\u012bva savu galveno m\u0113r\u0137i liel\u0101 m\u0113r\u0101 sasniedza un piev\u0113rsa pasaules uzman\u012bbu MI dro\u0161\u012bbas tematikai.<\/p>\n Eksperti br\u012bdina par m\u0101ksl\u012bg\u0101 intelekta riskiem<\/p>\n FLI atkl\u0101t\u0101 v\u0113stule nav iz\u0146\u0113muma gad\u012bjums, un aizvad\u012btajos gados daudzi zin\u0101tnieki, tehnolo\u0123iju l\u012bderi un dom\u0101t\u0101ji publiski br\u012bdin\u0101ju\u0161i par nekontrol\u0113tas MI att\u012bst\u012bbas draudiem. Piem\u0113ram, pla\u0161i paz\u012bstamais v\u0113sturnieks Juvals Noa Harari uzsver, ka \u0161\u012b tehnolo\u0123isk\u0101 revol\u016bcija var izr\u0101d\u012bties b\u016btiski at\u0161\u0137ir\u012bga no vis\u0101m iepriek\u0161\u0113j\u0101m un MI sp\u0113jas potenci\u0101li var k\u013c\u016bt tik varenas, ka riski civiliz\u0101cijai var\u0113tu b\u016bt pat eksistenci\u0101livii<\/a>. Harari un citi izce\u013c faktu, ka \u013coti att\u012bst\u012bts MI sp\u0113tu saprast cilv\u0113ku uzved\u012bbu un v\u0101j\u012bbas lab\u0101k nek\u0101 pa\u0161i cilv\u0113ki sevi, un t\u0101d\u0113j\u0101di var\u0113tu manipul\u0113t sabiedr\u012bbu l\u012bdz nepieredz\u0113tiem l\u012bme\u0146iem. \u0160\u0101ds MI, ja netiek pien\u0101c\u012bgi kontrol\u0113ts, var\u0113tu, piem\u0113ram, \u0123ener\u0113t \u0101rk\u0101rt\u012bgi p\u0101rliecino\u0161u dezinform\u0101ciju, radot draudus demokr\u0101tijai un sabiedriskajai k\u0101rt\u012bbai. Harari provokat\u012bvi jaut\u0101 – vai esam gatavi pasaulei, kur emocion\u0101li neitr\u0101ls superintelekts pie\u0146em l\u0113mumus cilv\u0113ku viet\u0101, turkl\u0101t l\u0113mumus, kuru motiv\u0101ciju pa\u0161i cilv\u0113ki vairs nesp\u0113j saprast?<\/p>\n Ar\u012b pa\u0161i MI tehnolo\u0123iju pionieri p\u0113d\u0113j\u0101 laik\u0101 izteiku\u0161ies p\u0101rsteidzo\u0161i atkl\u0101ti par \u0161iem riskiem. D\u017eefrijs Hintons, viens no neironu t\u012bklu izgudrot\u0101jiem un bie\u017ei d\u0113v\u0113ts par “MI krustt\u0113vu”, 2023. gad\u0101 pameta darbu Google<\/a>, lai br\u012bvi run\u0101tu par sav\u0101m ba\u017e\u0101m, un publiski l\u0113sa, ka past\u0101v aptuveni 10\u201320% varb\u016bt\u012bba, ka \u012bpa\u0161i att\u012bst\u012bts MI jau tuv\u0101ko tr\u012bs desmitga\u017eu laik\u0101 var\u0113tu izrais\u012bt cilv\u0113ces boj\u0101ejuviii<\/a>. \u0160\u0101ds skarbais nov\u0113rt\u0113jums apliecina, ka eksistenci\u0101lo apdraud\u0113jumu iesp\u0113jam\u012bbu apsver ar\u012b vado\u0161ie zin\u0101tnieki. L\u012bdz\u012bgi ar\u012b cits dzi\u013c\u0101s ma\u0161\u012bnm\u0101c\u012b\u0161an\u0101s pamatlic\u0113js, Tj\u016bringa balvas laure\u0101ts Jo\u0161u\u0101 Ben\u017eio skaidri atbalsta aicin\u0101jumu uz pagaidu “pauzi”, lai dotu laiku izstr\u0101d\u0101t dro\u0161\u012bbas pas\u0101kumus pirms m\u0113s speram n\u0101kamos so\u013cus arvien jaud\u012bg\u0101ka MI virzien\u0101ix<\/a>. \u0160o ekspertu nost\u0101jas nep\u0101rprotami liek saprast, ka MI att\u012bst\u012bba sola ne tikai ieguvumus, bet ar\u012b re\u0101lus draudus, ja netiks kontrol\u0113ta.<\/p>\n Protams, nevar aizmirst, ka MI att\u012bst\u012bbai ir ar\u012b milz\u012bgs pozit\u012bvais potenci\u0101ls no medic\u012bnas un zin\u0101tnes atkl\u0101jumiem l\u012bdz ikdienas dz\u012bves atvieglo\u0161anai. Tom\u0113r tie\u0161i sabalans\u0113ts skats uz iesp\u0113j\u0101m un riskiem b\u016bs kritiski svar\u012bgs, lai no tehnolo\u0123ij\u0101m g\u016btu labumu, vienlaikus izvairoties no nopietn\u0101m blakn\u0113m. K\u0101 nor\u0101da FLI p\u0101rst\u0101vji, sp\u0113c\u012bgu MI sist\u0113mu veido\u0161ana b\u016btu j\u0101turpina tikai tad, kad esam p\u0101rliecin\u0101ti, ka to efekts b\u016bs pozit\u012bvs un riski kontrol\u0113jami.<\/p>\n Ko tas noz\u012bm\u0113 Latvijai?<\/p>\n Latvijai glob\u0101laj\u0101 tehnolo\u0123iju sacens\u012bb\u0101 ir savi izaicin\u0101jumi un iesp\u0113jas. Lai ar\u012b MI “mil\u017eu c\u012b\u0146as” pamat\u0101 notiek lielvalstu un tehnolo\u0123iju korpor\u0101ciju starp\u0101, FLI v\u0113stul\u0113 paustie jaut\u0101jumi ir \u013coti b\u016btiski ar\u012b m\u016bsu kontekst\u0101. M\u0113s dz\u012bvojam digit\u0101li savienot\u0101 pasaul\u0113, un MI rad\u012bt\u0101s p\u0101rmai\u0146as un riski neapst\u0101jas pie valstu robe\u017e\u0101m. Dezinform\u0101cija internet\u0101, automatiz\u0101cijas ietekme uz darba tirgu, jaunu tehnolo\u0123iju rad\u012bt\u0101 ekonomisk\u0101 nevienl\u012bdz\u012bba – t\u0101s visas ir probl\u0113mas, kas var skart ar\u012b Latvijas sabiedr\u012bbu.<\/p>\n \u0160ie riski nav tikai teorija un, piem\u0113ram, v\u0113l\u0113\u0161anu ciklos tie jau par\u0101d\u0101s arvien konkr\u0113t\u0101k. Gaid\u0101mo Saeimas v\u0113l\u0113\u0161anu kontekst\u0101 MI var ietekm\u0113t v\u0113l\u0113t\u0101ju izv\u0113li gan ar maz\u0101k uzkr\u012bto\u0161iem krop\u013cojumiem, gan ar atkl\u0101t\u0101m manipul\u0101cij\u0101m. Piem\u0113ram, N\u012bderlandes datu aizsardz\u012bbas iest\u0101de br\u012bdin\u0101ja, ka v\u0113l\u0113\u0161anu laik\u0101 izmantotie \u010datboti sniedz neuzticamus un aizspriedumainus padomus, nereti virzot lietot\u0101jus uz atsevi\u0161\u0137\u0101m partij\u0101m neatkar\u012bgi no uzdotajiem jaut\u0101jumiemx<\/a>. \u0160\u0101di notikumi var graut uztic\u0113\u0161anos balso\u0161anas procesam un var tikt izmantoti ar\u012b Latvij\u0101, ja netiek noteikti skaidri kvalit\u0101tes un caurskat\u0101m\u012bbas standarti. <\/p>\n Citi precedenti r\u0101da klaju manipul\u0101ciju un m\u0113r\u0137tiec\u012bgas dezinform\u0101cijas kampa\u0146as. 2024. gad\u0101 \u0145\u016bhemp\u0161\u012br\u0101 tika izs\u016bt\u012bti ar MI balss klon\u0113\u0161anu rad\u012bti robotzvani, kas atdarin\u0101ja prezidenta D\u017eo Baidena balsi un attur\u0113ja pilso\u0146us no balso\u0161anas, un par to ASV regulators piem\u0113roja miljonu soda naudas, k\u0101 ar\u012b noteica aizliegumu \u0161\u0101d\u0101m AI robotzvanu bals\u012bmxi<\/a>. \u0160\u0101di piem\u0113ri skaidri nor\u0101da, ka tehnolo\u0123iski viegli reproduc\u0113jamas manipul\u0101cijas sp\u0113j m\u0113r\u0137\u0113ti ietekm\u0113t v\u0113l\u0113t\u0101ju uzved\u012bbu.<\/p>\n T\u0101p\u0113c Latvijai, t\u0101pat k\u0101 cit\u0101m Eiropas valst\u012bm, ir r\u016bp\u012bgi j\u0101seko MI att\u012bst\u012bbai un j\u0101iesaist\u0101s starptautiskos centienos nodro\u0161in\u0101t \u0161\u012bs jomas dro\u0161\u012bbu. Piem\u0113ram, Eiropas Savien\u012bba ir pie\u0146\u0113musi AI regulu, kas klasific\u0113 MI lietojumus p\u0113c riska l\u012bme\u0146iem un uzliek stingr\u0101kas pras\u012bbas augsta riska MI sist\u0113m\u0101m. Ar\u012b Latvijai ir j\u0101ievie\u0161 un j\u0101uzrauga \u0161\u012bs normas, lai pasarg\u0101tu savu sabiedr\u012bbu un lietot\u0101jus. Liel\u0101kais izaicin\u0101jums ir sabalans\u0113t tehnolo\u0123isko att\u012bst\u012bbu no biznesa perspekt\u012bvas, lai m\u016bsu uz\u0146\u0113mumi nezaud\u0113tu konkurences c\u012b\u0146\u0101, ar sabiedr\u012bbas interes\u0113m un dro\u0161\u012bbas jaut\u0101jumiem.<\/p>\n Balstoties uz Valsts kontroles 2025. gada situ\u0101cijas izp\u0113ti, valsts p\u0101rvald\u0113 MI izmanto\u0161ana strauji ien\u0101k, ta\u010du kop\u0113j\u0101 aina ir sadrumstalotaxii<\/a>. \u0160obr\u012bd 17 % iest\u0101\u017eu MI jau lieto, 22 % pl\u0101no ieviest, bet 55 % v\u0113l nav skaidra pl\u0101na. Bie\u017e\u0101kie pielietojumi ir pla\u0161i pieejamie \u0123enerat\u012bv\u0101 MI r\u012bki un virtu\u0101lie asistenti, ta\u010du tr\u016bkst vienotas uzskaites un krit\u0113riju, kas visp\u0101r kvalific\u0113jams k\u0101 MI valsts p\u0101rvald\u0113. Par b\u016btisk\u0101kajiem \u0161\u0137\u0113r\u0161\u013ciem iest\u0101des min finans\u0113jumu, datu aizsardz\u012bbu\/kiberdro\u0161\u012bbu un vadl\u012bniju tr\u016bkumu. 69 % iest\u0101\u017eu iek\u0161\u0113j\u0101s MI lieto\u0161anas vadl\u012bnijas v\u0113l nav izstr\u0101d\u0101tas. VARAM ir noteikta par vado\u0161o iest\u0101di, savuk\u0101rt 2025. gada 2. apr\u012bl\u012b dibin\u0101tais MI centrs form\u0101li sa\u0146\u0113mis uzdevumu sniegt konsult\u0101cijas un gatavot vadl\u012bnijas, tom\u0113r vienots meh\u0101nisms projektu uzskaitei un kopainas p\u0101rvald\u012bbai v\u0113l ir proces\u0101. <\/p>\n Normat\u012bv\u0101 un institucion\u0101l\u0101 vide tikai nostiprin\u0101s un, lai ieviestu ES AI akta pras\u012bbas, ir j\u0101groza virkne ties\u012bbu akti. Savuk\u0101rt ikgad\u0113j\u0101s \u012bsteno\u0161anas izmaksas Latvij\u0101 l\u0113stas ap 1 milj. euro. Vienlaikus ir pieejami ar\u012b resursi produktivit\u0101tes k\u0101pin\u0101\u0161anai, bet jau 2020. gad\u0101 tika iez\u012bm\u0113ts m\u0113r\u0137is l\u012bdz 2029. gadam sasniegt ~25 milj. euro gad\u0101 MI att\u012bst\u012bbai valsts sektor\u0101. Valsts kontrole zi\u0146ojum\u0101 uzsver nepiecie\u0161am\u012bbu no\u0161\u0137irt MI no visp\u0101r\u0113jas digitaliz\u0101cijas, ieviest vienotu uzskaiti un r\u0101d\u012bt\u0101jus, k\u0101 ar\u012b pabeigt nacion\u0101lo MI un datu strat\u0113\u0123iju, pret\u0113j\u0101 gad\u012bjum\u0101 gr\u016bti nov\u0113rt\u0113t ieguvumus, riskus un progresu. Praktiski tas noz\u012bm\u0113, ka j\u0101b\u016bt vienot\u0101m vadl\u012bnij\u0101m un risku p\u0101rvald\u012bbai, centraliz\u0113tai projektu uzskaitei un kvalitat\u012bvu datu pieejam\u012bbai.<\/p>\n Latvijas zin\u0101tnes un tehnolo\u0123iju kopiena var g\u016bt labumu no starptautisk\u0101s pieredzes gan m\u0101coties no lielo valstu k\u013c\u016bd\u0101m, gan iesaistoties kop\u012bgos p\u0113t\u012bjumos par MI dro\u0161\u012bbu. Mums ir sal\u012bdzino\u0161i neliela, ta\u010du akt\u012bva MI izstr\u0101d\u0101t\u0101ju ekosist\u0113ma, kas specializ\u0113jas konkr\u0113t\u0101s jom\u0101s, piem\u0113ram, dabisk\u0101s valodas apstr\u0101d\u0113 latvie\u0161u valod\u0101, finan\u0161u tehnolo\u0123ij\u0101s un citur. \u0160\u012bs ni\u0161as var k\u013c\u016bt par m\u016bsu priek\u0161roc\u012bbu, ja m\u0113s sp\u0113sim t\u0101s att\u012bst\u012bt, iev\u0113rojot augstus kvalit\u0101tes un dro\u0161\u012bbas standartus. Citiem v\u0101rdiem, Latvija var b\u016bt inovat\u012bva un konkur\u0113tsp\u0113j\u012bga MI jom\u0101, ta\u010du tikai tad, ja Latvijas sabiedr\u012bba un industrija uztic\u0113sies MI risin\u0101jumiem. Uztic\u012bba savuk\u0101rt rodas no p\u0101rliec\u012bbas, ka jaun\u0101s tehnolo\u0123ijas ir p\u0101rdom\u0101tas, p\u0101rbaud\u012btas un kalpo cilv\u0113ku interes\u0113m. No MI nav j\u0101baid\u0101s, bet j\u0101b\u016bt uzman\u012bgiem. Priek\u0161\u0101 ir ar\u012b liels skaidrojo\u0161ais un informat\u012bvais darbs, lai Latvijas sabiedr\u012bb\u0101, kuras digit\u0101l\u0101s prasmes par apm\u0113ram 10% joproj\u0101m atpaliek no ES vid\u0113j\u0101 l\u012bme\u0146axiii<\/a>, veicin\u0101tu uztic\u0113\u0161anos MI sniegtaj\u0101m iesp\u0113j\u0101m.<\/p>\n FLI pazi\u0146ojums par pauzi MI att\u012bst\u012bb\u0101 atg\u0101dina visiem, ar\u012b mums Latvij\u0101, ka tehnolo\u0123iju progress nav pa\u0161m\u0113r\u0137is. Ir j\u0101uzdod gr\u016bti jaut\u0101jumi: “K\u0101p\u0113c un k\u0101 m\u0113s att\u012bst\u0101m MI? Vai sp\u0113jam kontrol\u0113t t\u0101 sekas?”. \u0160os jaut\u0101jumus uzdod vado\u0161ie pasaules pr\u0101ti, un to vajadz\u0113tu dar\u012bt ar\u012b mums. Latvijai, veidojot savu digit\u0101lo n\u0101kotni, ir j\u0101saglab\u0101 l\u012bdzsvars starp atv\u0113rt\u012bbu inov\u0101cij\u0101m un piesardz\u012bbu pret riskiem. Tikai t\u0101 m\u0113s var\u0113sim pilnv\u0113rt\u012bgi izmantot m\u0101ksl\u012bg\u0101 intelekta sniegt\u0101s priek\u0161roc\u012bbas, vienlaikus pasarg\u0101jot savu sabiedr\u012bbu no tiem \u0113nas aspektiem, par kuriem br\u012bdina FLI un citi eksperti.<\/p>\n i<\/a> https:\/\/futureoflife.org\/open-letter\/pause-giant-ai-experiments\/<\/p>\n ii<\/a> https:\/\/www.reuters.com\/business\/finance\/opinions-split-over-ai-bubble-after-billions-invested-2025-10-16\/?utm_source=chatgpt.com<\/p>\n iii<\/a> https:\/\/finance.yahoo.com\/news\/first-big-winners-race-create-110000917.html?guccounter=1&guce_referrer=aHR0cHM6Ly93d3cuZ29vZ2xlLmNvbS8&guce_referrer_sig=AQAAAEFx1Uw6qePJi51LvUW175hyEetY5GibXNmnEuKOTg0h33xD1T57N7vMOy7wF8Z5_biD2F6YdhP6skLtnALmBLpNAe2gzAlaVA9O2YxgzjqQk4_Kdfg5KAVeiJJmCtCZx3U7JwTSvIYZ7Bp9aYWnb-O3Ej5YPD3N0-rzQxNX8h4E<\/p>\n iv<\/a> https:\/\/finance.yahoo.com\/news\/zuckerberg-is-pouring-billions-into-plans-for-personal-superintelligence-193923036.html?utm_source=chatgpt.com<\/p>\n v<\/a> https:\/\/www.mapfreglobalrisks.com\/en\/risks-insurance-management\/article\/race-ai-who-will-lead-future-superintelligence\/<\/p>\n vi<\/a> https:\/\/spectrum.ieee.org\/ai-pause-letter-stokes-fear<\/p>\n vii<\/a> Harari, Y. N. (2024). Nexus: A Brief History of Information Networks from the Stone Age to AI. Random House.<\/p>\n viii<\/a> https:\/\/www.theguardian.com\/technology\/2024\/dec\/27\/godfather-of-ai-raises-odds-of-the-technology-wiping-out-humanity-over-next-30-years#:~:text=Prof%20Geoffrey%20Hinton%2C%20who%20this,within%20the%20next%20three%20decades<\/p>\n ix<\/a> https:\/\/www.vice.com\/en\/article\/the-open-letter-to-stop-dangerous-ai-race-is-a-huge-mess\/<\/p>\n x<\/a> https:\/\/www.theguardian.com\/world\/2025\/oct\/21\/ai-chatbots-unreliable-biased-advice-voters-dutch-watchdog<\/p>\n xi<\/a> https:\/\/apnews.com\/article\/biden-robocalls-ai-new-hampshire-charges-fines-9e9cc63a71eb9c78b9bb0d1ec2aa6e9c<\/p>\n xii<\/a> https:\/\/www.lrvk.gov.lv\/lv\/getrevisionfile\/29802-kdDeizr20XEB2uoEZrVd5HxISMQ1zFcr.pdf<\/p>\n xiii<\/a> https:\/\/digital-strategy.ec.europa.eu\/en\/factpages\/latvia-2024-digital-decade-country-report?utm_source=chatgpt.com<\/p>\n Seko “Delfi” ar\u012b Instagram<\/a> un YouTube<\/a> profil\u0101 \u2013 pievienojies, lai uzzin\u0101tu svar\u012bg\u0101ko un interesant\u0101ko pirmais!<\/p>\n Publik\u0101cijas saturs vai t\u0101s jebk\u0101da apjoma da\u013ca ir aizsarg\u0101ts autorties\u012bbu objekts Autorties\u012bbu likuma izpratn\u0113, un t\u0101 izmanto\u0161ana bez izdev\u0113ja at\u013caujas ir aizliegta. Vair\u0101k lasi \u0161eit<\/a>