經濟合作與發(fā)展組織

譯題 2030年教育與技能的未來：概念性學習框架（一）

Product by OECD（Organization for Economic Co-operation and Development）

History and background of AI in education

Review of existing evidence

The development and application of Artificial Intelligence（AI） is bringing imminent and rapid change to almost every aspect of life， with todays children experiencing a very different life to that of their parents. To prepare people for the anticipated changes to their lives， we must ensure that our education and training is tuned to the new demands of the workplace and society. The landscape we must navigate is likely to be bumpy and there will be significant challenges， not least， those relating to ethics. Fundamental to success will be unpacking our relationship with the concept of Intelligence In thinking about how AI will impact on education and what sorts of knowledge and skills future citizens will need， we therefore need to look beyond the current trends towards trying to identify the jobs and skills that the world will require to the core issue of what it means to be intelligent in an AI augmented world. However， there is value in synthesizing what experts have investigated with respect to how susceptible jobs are to computerisation， because this is an important element of the context within which we need to reconceptualise human intelligence.

A seminal report from Frey & Osborne（2013） examined 702 detailed occupations， using Machine Learning AI， in the form of a Gaussian process classifier and found that 47 percent of total US employment is at risk and that wages and educational attainment exhibit a strong negative relationship with an occupations probability of computerisation.

In a second report in 2017， in the same authors concluded that： “In short， our findings suggest that recent development in Machine Learning will put a substantial share of employment， across a wide range of occupations， at risk in the near future. According to our estimates， however， this wave of automation will be followed by a subsequent slowdown in computers for labour substitution， due to persisting inhibiting engineering bottlenecks to computerisation.”

The bottlenecks are discussed in detail：1.Limitations of mobile robotics on perceptual and manipulation tasks. 2.Creative intelligence tasks which AI and machine learning cannot currently achieve. 3.Social intelligence tasks （the challenge of real time recognition of human emotion and how to respond intelligently to these.）

However， Fadel（2014） from a roundtable of experts made 6 predictions which ?provide some information about the sorts of jobs that may increase in the future：1.Routine tasks will remain the most automatable， but some facets of innovation and creativity may be automatable. 2. Complete adoption of technologies generally takes longer than anticipated but may be deeper than first assumed.3. Robust occupations are those with challenges， new discoveries， new performances and new things to be learnt and shared. 4.T shaped occupations， requiring both depth and breadth will see an increase in demand. 5.A top down review will not be able to predict future job patterns. This will have to come from sector by sector analysis. 6.There are many and variable parameters which interact with one another which need to be ?considered in order to predict future jobs.

譯文

人工智能在教育中的歷史和背景

審視現有證據

人工智能的發(fā)展和應用正在給生活多方面帶來快速的變化，今天的孩子們與父母有著非常不同的生活經歷。為了讓人們?yōu)槲磥眍A期的生活變化做好準備，我們必須確保我們所受的教育和培訓能夠適應工作場所和社會的新需求。未來我們必須應對的環(huán)境可能會崎嶇不平，而且將面臨重大挑戰(zhàn)，尤其是與倫理相關的挑戰(zhàn)。成功的基礎條件將揭示我們與智能概念的關系。在思考人工智能將如何影響教育和未來公民需要什么樣的知識和技能時，我們需要超越當前趨勢，試圖確定未來世界需要的工作和技能，即在一個人工智能增強的世界中被稱為智能意味著什么。然而，綜合專家們所調查的關于工作如何易受計算機影響的內容是有價值的，因為這是我們需要重新理解人類智能背景下的一個重要元素。

弗雷和奧斯本的開創(chuàng)性報告（2013）使用機器學習人工智能，以高斯過程分類器的形式，調查了702個詳細的職業(yè)，發(fā)現47%的美國就業(yè)風險與工資、教育程度呈現出強烈的負相關關系以及職業(yè)計算機化的概率。

在2017年的第二份報告中，同一作者得出結論：“簡而言之，我們的研究結果表明，機器學習的發(fā)展在不久的將來將使各種職業(yè)面臨大量就業(yè)風險。然而，根據我們的估計，由于計算機化的工程瓶頸，計算機的勞動力替代將會放緩。”

本文詳細討論了這些瓶頸問題：1.移動機器人在感知和操作任務上的局限性。2.人工智能和機器學習目前無法實現創(chuàng)造性智能任務。3.社會智力任務（人類情緒實時識別的挑戰(zhàn)以及如何智能地應對這些任務）。

然而，Fadel（2014）在一個專家圓桌會議上做出了6個預測，其中提供了一些關于未來可能增加的工作類型的信息：1.常規(guī)工作仍將是最可自動化的，一些需要創(chuàng)新和創(chuàng)造力的工作可能是可自動化的。2.完全采用技術通常比預期的要長，但可能比最初假設得更深。3.強大的職業(yè)是指那些有挑戰(zhàn)、有新發(fā)現和需要學習和分享的職業(yè)。4.需要深度和廣度的T型職業(yè)需求將會增加。5.自上而下的評估將無法預測未來的工作模式。這必須來自逐個部門的分析。6.為了預測未來的工作，需要考慮許多相互作用的可變參數。