Amangu ago: essuubi ly’obukodyo mu kutunda kwa gavumenti

Eby’obutale bya gavumenti eby’okutunda byajja bikozesebwa mu ngeri z’enjawulo olw’okufuna obuvunaanyizibwa, okwongera obusobozi, n’okufuna obuwanguzi mu kusasula. Ku biseera ebyogerwa, okutuusa mu myaka egy’omu maaso, eby’obutale byakozesebwa mu ngeri za makani zitundu; wabula kati obukozesebwa bwa algorithms n’obukozesebwa bwa software bitandise okuggyamu amaanyi mu kusalawo abasaala oba abalina akatale. Omulimu guno gutandise okusaba okusomesa amateeka agali mu maaso ku nteekateeka ezikola obulungi, okukuuma obutuufu, n’okuzuula bwekulana mu kusalawo.

Amateeka n’eby’enjawulo ebyakulabirirwa

Amateeka g’okusaba n’obuyinza mu kutunda aga gavumenti galina ebiseera eby’obukulu: okusooka okubuuza ensonga z’eby’entegeka n’emikisa, okugatta enteekateeka z’ebuzito n’obusobozi bw’abalondawo. Mu nsi yonna, okuyimiriza e-procurement kwaleeta okutumbula okw’enjawulo, okusobola okuteeka ku lusozi obutebenkevu n’okufuna amaanyi. Ekizikiza eky’obugumu kyajja wakati nga amawulire ag’okukola ga algorithm galina amagezi agakuliddwa. Mu 2019, ebitongole eby’enjawulo nga OECD byayise mu mateeka ga AI n’empisa, era mu maka ag’enjawulo abantu baatandika obukodyo okwetegereza emirimu gy’ebikozesebwa by’algorithms mu by’obuyinza.

Obukodyo obufuuse amateeka: ebigendererwa eby’obuyinza

Obukodyo obukoma mu nsi yonna buyise mu kukola amateeka agakyusa n’okuteeka enzikkiza ezikwata ku AI n’eby’algorithm. Ebyalo bingi bye byeyongera okwekenneenya: European Union yazziza omukwano ku nateekateeka eyitibwa AI Act (ekikwata ku mateeka agalambikiddwa okuva mu 2021 n’okubuulirwa mu 2023), era ebikolebwa n’obuvunaanyizibwa bw’eby’algorithm bisobola okuteeka ensonga ez’enjawulo mu kutunda. OECD ne bweyongera okubunyisa empisa z’okukozesa AI mu myaka egy’emirundi, okuyigiriza nti amawulire agawoomerwako era obutebenkevu mu by’obuyinza bisobola okulongoosa obulungi. Mu ngeri y’obusobozi, ebiwandiiko eby’enjawulo bisaba okuweebwa obulungi ku byalo by’okusaba n’okuzzaamu eby’obulamu bw’ebyaliggorithm.

Enfuna z’obufuzi n’ensonga mu nsi y’obulamu

Okugeza ku ngeri algorithm esobola okugezaamu kutunda, waliwo enfuuka z’eby’obulamu: okusasula obutebenkanye, okusala abalina obusobozi obutonotono mu kutunda, n’okuleetera okusonyiwa ku mbeera y’entambula y’obuwumbi. Era waliwo omukisa: okusobola okwongeza obutebenkevu mu kutuukiriza ebyetaagisa, okukyusa obutono obw’obuyinza mu kusala ebyetaago, n’okufuna ensonga z’okubala obulungi ku by’amaanyi ga ssente. Ebimu ku bintu ebyetaagisa okumanyiddwa nga byetaagisa okutuuka ku kigendererwa ky’obuvunaanyizibwa: human-in-the-loop, audit trails, transparency ya algorithmic models, n’empisa z’okulondoola amateeka mu kutunda.

Obuvunaanyizibwa bw’amateeka: ebyetaagisa n’ebiwandiiko

Okukakasa nti algorithms mu kutunda za gavumenti zikola mu ngeri etereeza obulungi kisaba amateeka ag’ensonga. Omuwendo gw’ebintu ebyetaagisa gulimu: okumanya ekitundu eky’obuzibu (algorithmic impact assessments), okuteeka amateeka g’okutereka obulakawo bw’ebyawandiiko n’obutebenkevu, okuwa omuntu akakasa ku nsonda z’amawulire, n’okukola empisa z’okukozesa obusobozi bw’obulungi. Ebitongole by’obuyambi byonna bisobola okuyiga mu nsi eno n’okuteeka ebikolebwa ebikakasa empisa. Ebitongole eby’enjawulo bya gavumenti bisobola okukola ebiragiro ebyetaagisa mu kutegeka amateeka ga procurement codes, nga bereeta enkola ey’okusoma algorithms era ne kugula abawandiisi abalikola.

Ebikolebwa n’okukulembera: emirimu egyetaagisa mu gavumenti

Okutereeza ekitongole ky’obuyinza mu kutunda okuva mu masanyalaze g’algorithms kityo kivaamu okukola ku mbeera ez’enjawulo. Ebikolebwa ebyetaagisa birimu: okuteeka ekibiina ekirabirwa obuyambi obukozesa technical audits, okuwandiika ensobi n’obulabirizi obw’obujjudde mu mabega ga procurement, okulondebwa kw’ebikozesebwa mu ngeri eno nga kiri ku mbeera y’amateeka, n’okukyusa amateeka ga procurement okusobola okutereka ensonga z’okukyusa algorithms. Okukuuma ekintu kino kulina okuwandiibwa mu mateeka, n’okutegekebwa okulongoosa abantu mu gavumenti abateeka mu bikorwa ebyo, wamu n’okubunyisa obuyambi obw’eby’obusobozi eri banabasaamu.

Obutongole n’eby’enjigiriza: enkola z’okuyamba n’obukuumi

Ekirala ekikulu kwe kujja n’emikisa gy’eby’enjigiriza mu nsi y’obulamu. Gavumenti zetaaga okugezaamu okusomesa abamu ku kulaba obulungi bwa algorithms, okutendereza obusasanyi mu kutunda, n’okukozesa empisa z’obulungi mu by’obusuubuzi. Okugeza, okutendereza eby’okusuubula ku nteekateeka y’omukozi ogw’enkola, okuteeka amasomo g’obukuumi bwa data mu by’obuyinza, n’okuteeka engeri z’obutebenkevu mu by’okukola ne bizibu. Okukola kino kisobola okuyamba ku kumanyisa abakozi n’abasaba ku mbeera y’okukola okutereeza obulungi mu kutunda kwa gavumenti.

Ebirina okukolebwa: empanvu n’obukakafu mu mateeka

Okumaliriza, enteekateeka ezikwatiddwa ku algorithms mu kutunda kwa gavumenti zisobola okuleetedde obulungi n’obuzibu. Abalondawo b’amateeka bagenda kwetaaga okwetegekera amateeka aganonyereza, okutuukiriza obulungi, n’okuteeka mu nkola empisa ezinza okukakasa obulungi. Eby’okukola birimu okutondawo audit plates, okukolera mu kussaamu amaanyi, n’okukakasa nti buli kimu kirina endowooza y’obuvunaanyizibwa. Bwe wabaawo ensonga z’obulabirizi, ekimala kye kiba kuteesa nti amateeka n’obukulembeze bikwatagana mu kusobola okukwata ku by’obulamu n’okuteeka ekirungi mu kutunda kwa gavumenti.