Ongo’i Mateaki’anga ʻo e Kau Tokoni ʻIkai Lava Ke Fakapapau’i i he Faleako Fakapo’uli
Ko e ongo’i mateaki’anga ko ia kuo tau kotoa fehi’a ai ko ha kau ako, ʻi he taimi ko ia ke fakalelei’i ai ha ngaahi essay ʻi he po, ʻo pehē. ʻOku ke lau ha’ana ke fakahaa’i ʻe ha tokotaha ako—e fakatātaaa’i, ko ha Akolea-ʻUtu mo e Lea Fakaliki—ʻa e lea ʻoku nau fili ai e ngaahi lea lahi mo poto mo e ngaahi sakinika fakamatalá ʻoku matu’aki fakapapau’i mo lōloa, ka ʻikai ia ke kei fe’unga mo a latou ngāue he kalasi fuoloa kuo nau fai. Ko e masiva vave taha ko e masiva ko e ngāue fakapuhiako (academic dishonesty), ka kapau te ke tamate’i e lea ʻaki e ngaahi siaki fakafananaʻanga (similarity checkers) tu’uma’u, ʻoku ʻikai ha taha ʻoku ʻilo’i. Ko e ha ʻeni: ko e kapau oku tokoni mai ʻa e kau akó mei ha ngaahi tokoni/kaunga te ke lava ʻoua ʻe lau? Ko e fakamahino ko ia: te nau tokoni atu ki ha ngaahi me’a tohi i ha lea kehe, ʻo liliu ia tonu ki he Lea faka-ʻEgelani. Ko e fakalukufua ki ai, ʻo e tupu vave lahi ʻa e generative AI, ʻoku tuʻu ai ʻa e kau faiaako lahi mo e ongo’i fakalongolongo mo e loto vaivai. Ko hono fakangofua ʻo e AI detection pē ke fakahoko ʻaki ʻi he’ene tu’uma’u kuo pau ke ke toe fakakaukau ke mo’ui mo fakapotopoto lelei, koe’uhi ko e tokoni ko ia ʻe taha pē ʻoku ʻikai toe lava ia ke ta’ofi e tapu fakasino’aki e ako (academic integrity) ʻi he ngaahi kalasi kehekehe mo e ngaahi tekinolosia fakalangi kuo lahi.
Fefanongonongo he Lea mo e Sipa’i AI ʻOku Tui’aki Masiofo
Ko e kalasi fakapo’uli faka ongo’i ʻaonga ko ia ʻo e kuonga nei ko ha feitu’u ola lahi mo e leó ni, ʻo nau si’i mei he ngaahi lea kehekehe—ko ia ʻoku a’u ai e ngaahi me’a faka-kemipea lahi mo e ngaahi faingata’a fakapa’anga pehe ki he mo’oni ʻa e tauhi a e kau akó. Kapau ʻoku fehangahangai ʻa e kau akó mo e fefanongonongo he lea ʻi he plagiarism, ʻe lava ke nau liliu ki he liliu mai ʻa e ngaahi tokoni/kaunga fakavahevahe ʻi he fonua kehe, ʻo fakavave’i fakalua ai e ngaahi siaki fakafananaʻanga tu’uma’u ia ʻoku si’i pē he ngaahi database ʻi he Lea faka-ʻEgelani. Foki, ko e fakabataní ʻo e generative AI ki he ngāue ʻa e kau akó kuo liliu ta’emalava e tafa’aki ʻo e fakapulipuli fakafakaako. ʻOku tau fehangahangai mo ha ongo me’a fakamata’u ʻe ua: ko e plagiarism kuó nau liliu ia (translated plagiarism) mo e tohi ʻoku poto mo faka’uli’i e masini (sophisticated machine-generated text).
ʻOku mahu’inga ke tau ʻilo’i e ngaahi fakangatangata fakatēkinikale ʻo e ngaahi me’angaue siaki AI ʻi he taimi ni. Ko e ngaahi sisitemi ko ia ʻoku ngāue’aki ki he ngaahi fakamahino fakasiteisi (statistical probabilities), ʻo liliu ai ki he ngaahi fakamatala hangē ko e perplexity mo e burstiness ke “tānaki” pe ko e tauhi a e tangata pe ko e tauhi a ha masini ʻoku fai e tohi. Kositupú, koe’uhi ko e probabilistic ia, ʻoku ha’ele ai e ngaahi palopalema lahi, ko e taha ko e false positives mo e false negatives. Ko e false positive—ko ia kapau ʻoku fakailifia’i fakamahu’inga’i ʻa e tohi lelei a e kau akó ko e “AI-generated” ka ʻikai—e lava ke ne faʻahinga’i fakalautaha e va fetu’utaki iá mo e faiaako mo e kau akó, pea tokoni ke tupu e tokanga lahi mo e mo’ua tupu loto he tokotaha ako. ʻOsi hono tukuange, ko e false negatives ʻoku tuku ai e fakapuhiako lahi ke ta’e iloa’i, ʻo hangē ko ha fuga’i. Ko e kau faiaako, kuo pau ke tau ʻiloʻi ko e ngaahi me’angaue siaki ʻoku ʻikai ko ha tu’utu’uni fakamutunga ia e mo’oni. Ko e ngaahi me’angaue ia ʻoku ʻikai lelei, pea ʻe ʻikai lava ia ke sui e ʻilo faka’amanaki mo e fakapoto ʻoku loaʻa ki he faiaako fekau’aki mo e mafai mo e tupu ʻa e kau akó.
Fakasi’isi’i Ako Ki he Fua’i Fakalelei tukupau (Process-Based Assessment) mo e Ako Fakamo’oni
ʻI he kaha’u, kuo tau liliu e taumu’a mei he siaki fakamata’ala ki he ngaahi fo’i tali fakalēlei (proactive) ʻi he tafa’aki ako. Ko e tali ki he ngaahi faingata’a ko ia ʻoku lolotonga, ʻoku tu’u ia ʻi he process-based assessment kae ʻikai ko e tauhi pē ki he fua fakamuimui. ʻO kapau te tau fakapapau’i e malaga pe “journey” ʻo e hiki tohi, ʻe lava ke tau langa e ʻilo’i ʻe he kau akó ʻoku nau lava (student self-efficacy) pea fakapapau’i e ako fakamo’oni ke hoko me’a ʻikai ke ta’emotu e tokanga’i kae pule’i pē e algorithms ʻoku mafu’i palopalema.
Ko e rautaki kamata taha ko hono fakaʻaonga’i ʻo e document version history ko ha konga angamaheni (standard component) ʻo e fakama’ala’ala/pule’i tohi. Ko e ngaahi pa’anga hangē ko Google Docs ʻoku ai e faingamālie ki he kau faiaako ke sio ki he processus kotoa ʻo e taau’ilo, pea ke sio ki he founga ʻoku langa ai ʻe ha tokotaha ako ʻa ʻene ngaahi palopalema fakapoto (arguments) ʻi hono taimi. Ko hono hū vave mai ʻa e ngaahi kupu lahi ʻo ha tohi lelei kuo ʻikai ke kamata ʻaki ʻi he taimi kuo te’eki ai ke tupi e ki (typing history) ʻe lava ke fakahaa’i ke tau masiva ki he translated plagiarism pe pe ko e generative AI. Ko e ngāue ko ia ʻoku liliu ai e talanoa mei he faka’ilo (accusation) ki ha talanoa felāve’i ʻoku kau ki he founga hiki tohi ia pē.
Ko e rautaki hono ua ʻoku kau ai e fiema’u ke fai e iterative drafting mo e formative assessment ʻoku hokohoko. Kapau ʻoku vahe e ngaahi assign ki ha ngaahi mílestone pē ʻoku lava ke pule’i—hangē ko e brainstorm, outline, drafting, mo e revising—e to’ua ai e kau akó mei he loto-mamahi ke nau “panikí” pea hū kehe ki he fakapuhiako fakafakaako. ʻO tuku atu e palaute ʻi he sitepu kotoa, ʻoku langa ai ha feitu’u fakatu’upake (scaffolded environment) ʻoku fehoanaki ai e faiaako mo e tupu’anga ʻo e manatu ʻa e tokotaha ako. Ko e founga ko ia ʻoku fakaafe’i ai ʻa e ngāue ko e hāmai e ngaahi tokoni/kaunga ʻikai fakapapau’i ʻo e fonua kehe pe ko e AI tools, koe’uhi he tokotaha ako kuo pau ke ne faʻa fakahaa’i pē e tupu ha ne ʻilo fakakaukau (evolving understanding).
Ko e rautaki hono tolu ko hono fokotu’u ʻo e prompts ʻoku matu’aki fakapitoa mo fakafounga ki ha context. Ko e ngaahi topic essay faka’aho (generic essay topics) ʻoku lava ke liliu vave ki generative AI pe lava ke ma’u ia mei ha ngaahi tohi kehe ʻoku lolotonga. ʻE tatau ke tau fokotu’u e ngaahi assignment ʻoku fiema’u ai e kau akó ke nau faka-kau ai e ngaahi manatu ʻo e kalasi ki he ʻenau ngaahi a’usia pē, ngaahi talanoa ʻo e kalasi talu ai, pe ha ngaahi me’a fakapa’anga pē (highly specific local events). Ko e fokotu’u assignment fakamo’oni ʻoku fakapapau’i ai e kau akó ke nau fakahoko ʻa e ngāue fakakaukau loloto ki he me’a, ʻo e’u atu ai ke toe faigata’a lahi ke nau faka’uli’i ke fakamavae mei he ngāue fakakaukau ʻoku fiema’u ke fai e tali tupu-tatau (original response).
Fetu’utaki ki he Kaha’u mo ha Fakautuutu pea mo e Poto Fakapolofesinale
Ko e faito’o ʻo e ako ʻoku ʻikai ke toe tu’utāmaki—ko e faingata’a ʻo e translated plagiarism mo e generative AI ʻoku kei hoko atu pē. ʻOku malava e ongo’i vave ke ʻeke ha tool siaki AI lelei taha, ka te tau lava ke tau fakapapau’i e integrity ʻaki e founga lahi: ʻoku tokoni mai e tu’unga fakatēkinikale mo e tafa’aki ako. ʻO kapau te tau tali e process-based assessment, fokotu’u e ngaahi ngāue fakamo’oni, pea mo ta’ofi e tokanga ki he tupu ʻa e kau akó, ʻe lava ke tau fakapapau’i e nofo ʻa e kalasi ʻi ha ngaahi feitu’u ʻo e ako fakamo’oni. Ko e kau faiaako, ko e me’angāue lahi taha ʻoku ʻikai ko ha algorithm, ka ko ʻetau poto fakapolofesinale pea mo ʻetau tu’utu’uni ke langa e fakamo’oni ʻa e kau akó. ʻOku tau ma’u e mafai ke fetu’utaki, ke tataki e kau akó, pea ke ola lelei ʻi he kuonga fo’ou ʻo e ako.
