Code Velocity
Utafiti wa AI

Kielelezo cha Ufasaha wa AI: Kupima Ustadi wa Ushirikiano wa Binadamu na AI

·7 dakika kusoma·Anthropic·Chanzo asili
Shiriki
Picha inayoonyesha dhana ya ufasaha wa AI na ushirikiano wa binadamu na AI, na data.

Ufasaha Kwanza: Kielelezo cha AI cha Anthropic kwa Ushirikiano Wenye Ustadi

Ujumuishaji wa haraka wa zana za AI katika shughuli za kila siku umekuwa wa kushangaza sana. Hata hivyo, kadri AI inavyozidi kuenea kila mahali, swali muhimu linatokea: je, watumiaji wanapokea tu zana hizi, au wanakuza ustadi muhimu wa kuzitumia kwa ufanisi? Anthropic, kiongozi katika ukuzaji wa AI inayowajibika, inalenga kujibu swali hili kwa Kielelezo chake kipya cha Ufasaha wa AI, ripoti mpya iliyoundwa kupima na kufuatilia mageuzi ya ustadi wa ushirikiano wa binadamu na AI.

Ripoti za Elimu za Anthropic za awali zilifichua jinsi wanafunzi wa vyuo vikuu na waelimishaji wanavyotumia mifumo ya hali ya juu kama Claude kwa kazi mbalimbali kuanzia kutengeneza ripoti hadi kupanga masomo. Hata hivyo, tafiti hizi ziligusa zaidi nini watumiaji walikuwa wakifanya. Kielelezo cha Ufasaha wa AI kinaingia ndani zaidi, kikichunguza jinsi watu binafsi wanavyoshirikiana na AI, kikianzisha mfumo wa kuelewa "ufasaha" na teknolojia hii yenye mabadiliko.

Kufafanua Ufasaha wa AI: Mfumo wa 4D

Ili kupima ufasaha wa AI, Anthropic ilishirikiana na Maprofesa Rick Dakan na Joseph Feller kuunda Mfumo wa Ufasaha wa AI wa 4D. Mfumo huu mpana unatambua tabia 24 maalum zinazoonyesha ushirikiano salama na wenye ufanisi wa binadamu na AI. Kwa madhumuni ya utafiti huu wa awali, Anthropic ililenga tabia 11 zinazoonekana moja kwa moja ndani ya kiolesura cha gumzo cha Claude.ai. Zingine 13, ambazo zinajumuisha vipengele muhimu kama vile kuwa mkweli kuhusu jukumu la AI kazini au kuzingatia matokeo ya matokeo yanayotokana na AI, hutokea nje ya gumzo na zitachunguzwa katika utafiti wa ubora wa baadaye.

Kwa kutumia zana ya uchambuzi inayolinda faragha, timu ya utafiti ilichunguza kwa makini mazungumzo 9,830 ya zamu nyingi kwenye Claude.ai kwa kipindi cha siku 7 mnamo Januari 2026. Data hii kubwa ilitoa msingi imara wa kupima uwepo au kutokuwepo kwa tabia 11 za ufasaha zinazoonekana, na kusababisha kuundwa kwa Kielelezo cha Ufasaha wa AI. Kielelezo hiki kinatoa picha ya mifumo ya ushirikiano ya sasa na msingi wa kufuatilia mageuzi yake kadri mifumo ya AI inavyoendelea.

Nguvu ya Kurudia na Kuboresha katika Mwingiliano wa AI

Moja ya matokeo yenye kushawishi zaidi kutoka Kielelezo cha Ufasaha wa AI ni uhusiano mkubwa kati ya kurudia na kuboresha na karibu tabia zote zingine za ufasaha wa AI. Utafiti ulifichua kuwa 85.7% ya mazungumzo yalihusisha watumiaji kujenga juu ya mwingiliano uliopita ili kuboresha kazi zao, badala ya kukubali tu jibu la awali. Mazungumzo haya ya kurudia yalionyesha viwango vya juu zaidi vya tabia zingine za ufasaha, yakiongeza maradufu umahiri unaoonekana katika mazungumzo ya haraka, ya kurudiana.

Athari za Kurudia kwenye Tabia za Ufasaha wa AI

Kiashiria cha TabiaMazungumzo yenye Kurudia & Kuboresha (n=8,424)Mazungumzo bila Kurudia & Kuboresha (n=1,406)Kipengele cha Ongezeko (Kurudia vs. Isiyo ya Kurudia)
Kuhoji Mantiki ya ClaudeJuuChini5.6x
Kutambua Muktadha UliopoteaJuuChini4x
Kufafanua LengoJuuKati~2x
Kutaja UmbizoJuuKati~2x
Kutoa MifanoJuuKati~2x
Wastani wa Tabia za Ziada za Ufasaha2.671.332x

Jedwali: Linaloonyesha ongezeko la kuenea kwa tabia za ufasaha katika mazungumzo yenye kurudia na kuboresha.

'Athari hii ya kurudia na kuboresha' inasisitiza umuhimu wa kuichukulia AI kama mshirika wa kufikiri badala ya mtekelezaji wa kazi tu. Watumiaji wanaoshiriki kikamilifu katika mazungumzo, wakipinga na kuboresha maswali yao, wana uwezekano mkubwa zaidi wa kutathmini kwa kina matokeo ya AI, kuhoji mantiki yake, na kutambua muktadha muhimu uliopotea. Hili linaendana na dhana ya mtiririko wa kazi unaoendeshwa na wakala, ambapo usimamizi wa binadamu na maoni ya kurudisha mara kwa mara huleta matokeo bora, kama ilivyochunguzwa katika majadiliano kuhusu mifumo kama Mifumo ya Kazi ya Wakala ya GitHub.

Upanga Wenye Makali Mawili wa Uundaji wa Visanii vya AI

Ingawa kurudia huongeza ufasaha kwa ujumla, ripoti ilifichua muundo wenye utata pale watumiaji wanapoielekeza AI kutengeneza visanii kama vile msimbo, nyaraka, au zana shirikishi. Mazungumzo haya, yanayowakilisha 12.3% ya sampuli, yalionyesha watumiaji wakizidi kuwa waelekezaji lakini cha kushangaza, hawathaminiki sana.

Wanapounda visanii, watumiaji walikuwa na uwezekano mkubwa zaidi wa kufafanua malengo yao (+14.7 asilimia pointi), kutaja fomati (+14.5pp), na kutoa mifano (+13.4pp). Hata hivyo, uelekezaji huu ulioongezeka haukusababisha uelewa mkubwa zaidi. Kwa kweli, watumiaji walikuwa na uwezekano mdogo zaidi wa kutambua muktadha uliopotea (-5.2pp), kuangalia ukweli (-3.7pp), au kuhoji mantiki ya mfumo (-3.1pp). Mwelekeo huu una wasiwasi mkubwa ikizingatiwa kuwa kazi ngumu, ambazo mara nyingi zinahusishwa na uundaji wa visanii, ndipo mifumo ya AI kama Claude Opus 4.6 au hata mifumo ya hali ya juu kama GPT-5 (kama ingekuwa inatumika, ingawa kiungo kinaelekeza kwenye toleo la baadaye au la kudhaniwa) zina uwezekano mkubwa wa kukutana na matatizo.

Hali hii inaweza kuhusishwa na matokeo yanayotokana na AI yanayoonekana kuwa yamekamilika na yenye utendaji mzuri, ambayo yanaweza kuwapumbaza watumiaji na kuwafanya waamini kuwa kazi imekamilika kabla ya wakati. Iwe ni kubuni kiolesura cha mtumiaji (UI) au kuandaa uchambuzi wa kisheria, uwezo wa kuchunguza kwa kina matokeo ya AI bado ni muhimu sana. Kadri mifumo ya AI inavyozidi kuwa ya kisasa, hatari ya kukubali bila kuchunguza matokeo yanayoonekana kuwa kamilifu huongezeka, hivyo kufanya ustadi wa kutathmini kuwa muhimu zaidi kuliko hapo awali.

Kukuza Ufasaha Wako wa AI

Habari njema ni kwamba ufasaha wa AI, kama ustadi wowote, unaweza kuendelezwa. Kulingana na matokeo yao, Anthropic inatoa ushauri wa vitendo kwa watumiaji wanaotaka kuboresha ushirikiano wao na AI:

  1. Kuendelea kwenye Mazungumzo: Kubali majibu ya awali ya AI kama sehemu ya kuanzia. Shiriki katika maswali ya kufuatilia, pinga mawazo, na uboreshe maombi yako mara kwa mara. Ushirikishwaji huu wa kazi ni kiashiria chenye nguvu zaidi cha tabia zingine za ufasaha.
  2. Kuhoji Matokeo Yaliyoboreshwa: Wakati mfumo wa AI unatoa kitu ambacho kinaonekana kimekamilika na sahihi, pumzika na utumie mawazo ya kina. Uliza: Je, hii ni sahihi kweli? Je, kuna chochote kinachokosekana? Je, mantiki inasimama? Usiruhusu urembo wa kuonekana uzidi tathmini muhimu.
  3. Kuweka Masharti ya Ushirikiano: Bainisha mapema jinsi unavyotaka AI iingiliane nawe. Maelekezo wazi kama 'Pingana kama mawazo yangu si sahihi,' 'Nionyeshe mantiki yako,' au 'Niambie unachokipata changamoto' yanaweza kubadilisha msingi wa mienendo, yakikuza ushirikiano ulio wazi na imara zaidi.

Msingi kwa Ukuaji wa Ustadi wa AI wa Baadaye

Ni muhimu kutambua mapungufu ya utafiti huu wa awali. Sampuli, inayojumuisha watumiaji wa Claude.ai wanaoshiriki katika mazungumzo ya zamu nyingi kutoka mapema mwaka 2026, inaweza kuelemea kwa watumiaji wa mapema ambao tayari wamezoea AI, na sio idadi kubwa ya watu. Utafiti pia unazingatia tu tabia zinazoonekana ndani ya kiolesura cha gumzo, ukiacha tabia muhimu za matumizi ya kimaadili na yanayowajibika zinazotokea nje. Tahadhari hizi zinamaanisha Kielelezo cha Ufasaha wa AI kinatoa msingi kwa kundi hili maalum la watu na sehemu ya kuanzia kwa utafiti wa kina zaidi, wa muda mrefu.

Licha ya mapungufu haya, Kielelezo cha Ufasaha wa AI kinaashiria hatua muhimu kuelekea kuelewa na kukuza ushirikiano wenye ufanisi kati ya binadamu na AI. Kadri zana za AI zinavyoendelea kukua, kuwawezesha watumiaji na ustadi wa kushiriki kwa umakini, kwa kurudia, na kwa kuwajibika kutakuwa muhimu katika kutambua uwezo kamili wa teknolojia hii huku ikipunguza hatari zake. Ripoti hii ya awali inaweka msingi kwa utafiti wa siku zijazo, ikiahidi kuwaongoza watumiaji na watengenezaji katika kujenga mustakabali wenye ufasaha zaidi na wenye manufaa unaotegemea AI.

Maswali Yanayoulizwa Mara kwa Mara

What is the Anthropic AI Fluency Index?
The Anthropic AI Fluency Index is a new metric developed by Anthropic to assess how well individuals are developing skills to effectively use AI tools. Moving beyond mere adoption, the index tracks 11 directly observable behaviors that represent safe and effective human-AI collaboration, based on the 4D AI Fluency Framework. It aims to provide a baseline measurement of user proficiency, helping to understand how these critical skills evolve as AI technology becomes more integrated into daily life. The initial study analyzed nearly 10,000 conversations on Claude.ai to identify key patterns in user interaction and skill development.
How is AI fluency measured by Anthropic?
AI fluency is measured by tracking the presence or absence of 11 specific behavioral indicators during user interactions with Claude on Claude.ai. These indicators are derived from the broader 4D AI Fluency Framework, which defines 24 behaviors of safe and effective human-AI collaboration. For the initial study, Anthropic utilized a privacy-preserving analysis tool to examine 9,830 multi-turn conversations over a 7-day period. Behaviors like 'iteration and refinement,' 'questioning reasoning,' and 'identifying missing context' were observed and classified as present or absent within each conversation, providing a quantitative baseline for AI proficiency.
What is the 'iteration and refinement effect' in AI fluency?
The 'iteration and refinement effect' refers to the strong correlation found between users who build on previous exchanges to refine their work with AI, and the display of other key AI fluency behaviors. Conversations exhibiting iteration and refinement—meaning users don't just accept the first AI response but actively engage in follow-up questions, pushbacks, and adjustments—showed significantly higher rates of other fluency indicators. For instance, these iterative conversations were 5.6 times more likely to involve users questioning Claude's reasoning and 4 times more likely to identify missing context, underscoring the importance of sustained, dynamic engagement for developing AI proficiency.
Why do users become less evaluative when creating artifacts with AI?
Anthropic's research found that when users engage AI to create artifacts such as code, documents, or interactive tools, they tend to become more directive but paradoxically less evaluative. This means users are more likely to clarify goals and provide examples, but less likely to question the model's reasoning, identify missing context, or check facts. Possible explanations include the polished appearance of AI-generated outputs, which might lead users to prematurely trust the results, or the nature of certain tasks where functional aesthetics might outweigh factual precision. Regardless, this pattern highlights a critical area for improvement in human-AI collaboration, emphasizing the need for continued critical assessment even with seemingly complete outputs.
How can individuals improve their AI fluency according to Anthropic?
Anthropic suggests three key areas for individuals to enhance their AI fluency. First, 'staying in the conversation' means treating initial AI responses as starting points, asking follow-up questions, and actively refining outputs. Second, 'questioning polished outputs' involves critically evaluating AI-generated artifacts for accuracy, completeness, and logical soundness, even if they appear perfect. Third, 'setting the terms of the collaboration' encourages users to explicitly instruct AI on how to interact, for example, by asking it to explain its reasoning or push back on assumptions. These practices aim to foster deeper engagement and critical thinking in human-AI interactions.
What are the limitations of the AI Fluency Index study?
The initial AI Fluency Index study has several important limitations. The sample is restricted to Claude.ai users engaging in multi-turn conversations during a single week in January 2026, which likely skews towards early adopters and may not represent the broader population. The study also only assesses 11 out of 24 behaviors from the 4D AI Fluency Framework, focusing solely on directly observable interactions within the chat interface, thus missing crucial ethical and responsible use behaviors that occur externally. Furthermore, the binary classification of behaviors might overlook nuanced demonstrations, and it cannot account for 'implicit behaviors' where users might mentally evaluate AI outputs without verbalizing their critical assessment in the chat.

Baki na Habari

Pokea habari za hivi karibuni za AI kwenye barua pepe yako.

Shiriki