AI ba sabon abu ba ne gaba ɗaya — an fara tunaninsa tun 1950s. Amma sai kwanan nan ya bunƙasa har ya zama batun kowa. Ga taƙaitaccen tarihinsa.
Farko (1950s-1990s)
An fara tunanin AI a hukumance a taron Dartmouth na 1956. A shekarun farko, AI ya kasance galibi tunani da bincike, ba tare da babban amfani na aikace ba, saboda ƙarancin bayanai da ƙarfin kwamfuta.
Babban ci gaba (2000s-2020s)
Haɗuwar bayanai masu yawa (internet), ƙarfin kwamfuta, da deep learning ya canza komai. Manyan model na harshe (LLM) suka bullo, har ChatGPT (2022) ya kawo AI ga kowa.
Zamanin yanzu (2023-2026)
Yanzu zamanin AI agent ne — AI da ke yin ayyuka da kansa. Amma Hausa da sauran harsuna masu ƙarancin bayanai har yanzu suna baya. Babu cikakken LLM na Hausa mai amfani a sikeli ya zuwa 2026. Mafi kusa shi ne N-ATLaS (Awarri da NITDA/NCAIR, 2025) — gyaran Llama-3 8B ne ƙarƙashin lasisi mai iyakance amfani ga kasuwanci har wajen masu amfani 1,000; InkubaLM kuma na Lelapa, ba na kasuwanci ba ne (CC BY-NC). Saboda haka ka ɗauke su a matsayin fasaha mai bunƙasa, ba wadda ta cika ba.
Idan kana tunanin yin amfani da AI a kasuwanci, maimakon kayan aiki da yawa daban-daban, akwai dandamali da ke haɗa ayyukan AI (tattaunawa, sarrafa kai, da app) wuri ɗaya — kamar osFoundry, dandamalin AI mai aiki da kansa (agentic).
Karin karatu
Wannan bayani ne na gamagari, ba shawarar ƙwararru ba.