《教育中的人工智能：學(xué)習(xí)速度的變化》（五）

Product by UNESCO IITE （The UNESCO Institute for Information Technologies in Education）聯(lián)合國(guó)教科文組織教育信息技術(shù)研究所

該報(bào)告為UNESCO IITE新推出的“教育的數(shù)字化轉(zhuǎn)型”（Digital Transformation of Education）系列出版物的第一期。據(jù)悉，“教育的數(shù)字化轉(zhuǎn)型”系列將包括政策簡(jiǎn)報(bào)、分析報(bào)告和反思論文，以探討由于技術(shù)的使用及其對(duì)教育和人類生活其他領(lǐng)域的影響，教育領(lǐng)域正在發(fā)生和正在出現(xiàn)的根本性變化。

4. Challenges

4.1 Equity， Equality and Access

The term ‘digital divide has been in use since the late 1990s. Although the manual recording of data by educators and institutions can help to address this divide， as electronic data capture becomes the norm， learners who have limited access to the information required to assist them in their studies， and the ability to generate and share electronic data， may be disadvantaged when it comes to the building of personalized learning pathways which adequately identify and address their needs.

4.2 The Ethical Dimension

At present it is private companies who are leading the way in bringing the power of AI to education， giving rise to concerns about the privacy， protection and use of student data. An ethical foundation for AI is required during the development of AI-enabled services and solutions： not just at the deployment stage. How can we ensure that gender， racial， socio-economic and ability biases are not introduced at the programming level？ How can we ensure that social and cultural stereotypes are not promulgated？ How can we ensure that all learners， regardless of where they live， have the same access to the benefits which will accrue？

4.3 Technology Dependence

An increased reliance upon AI will not be entirely to our benefit. We can expect to see an increased number of jobs replaced by automation， in both the developed and developing world. There is also a danger that our ability to delegate tasks and cognitive functions to machines can increase our dependence upon technology whilst eroding our own ability to perform these， in what might be termed a ‘use it or lose it scenario. Do students who rely on a computer keyboard lose the ability to write legibly？ Do those who perform calculations using a spreadsheet or calculator become less skilled at mental arithmetic？ Has a reliance upon GPS affected our ability to navigate？ As these technologies become smarter and more capable， it will also become more important than ever to know where it is we want to go.

4.4 Continuous Professional Development

To derive the benefits of AI in education， it will be necessary to equip educators and administrators with the skills to assess and interpret the results. Effective AI solutions generate easy-to-interpret visualizations and dashboards which allow users to interrogate the data and absorb insights in real time. Custom views and reports， tailored for different roles and functions， can easily be produced. The core focus of CPD in this context involves training educators to leverage the insights generated so as to incorporate them into their teaching practice， and to assist them in creating personalised learning pathways for their students.

5. Whats Next？

5.1 New Ways of Doing

5.1.1 Metacognitive Scaffolding

AI can play a role in metacognitive scaffolding by allowing the students themselves to benefit from the insights generated during the course of their learning pathways. Increasingly， students will become the primary users of AI solutions and services， rather than simply the subjects of the data analyzed by educators， administrators and system owners. A learner who is provided with greater insight into how they learn， and how they think， has greater agency and control over their education， and is equipped with vital knowledge of self to act as a bedrock for lifelong learning.

5.1.2 Personalized Assessment and Credentialing

AI will also play an important part in addressing the next great challenge being offered to education technologists： how to support personalized assessment. We know that current forms of assessment in our schools and colleges are seldom aligned to the skills that will be demanded of students when they enter the world of work. Some of the higher-order thinking skills that will be required in a 21st century environment — recall， comparison， analysis and inference — but soft skills， ‘people skills， moral character， teamwork， collaboration and the ability to work effectively as part of a team are difficult to evaluate using these traditional forms.

Many institutions and an increasing number of school systems are investigating the use of micro-credentials that measure complex and general learning capabilities， including such 21st century skills as critical thinking， creativity， communication skills and entrepreneurship. If these new credentials are to remain relevant after their introduction it will also be essential to conduct a continuous analysis of their efficacy and effectiveness， and to present these findings and insights using real-time dashboards tailored to each stakeholders area of interest or expertise.

譯文

4. 挑戰(zhàn)

4.1 平等、公平與機(jī)會(huì)

“數(shù)字鴻溝”一詞自20世紀(jì)90年代末開始使用。盡管教育工作者和機(jī)構(gòu)手工記錄數(shù)據(jù)有助于彌補(bǔ)信息富有者與信息貧困者之間的差距，但隨著電子數(shù)據(jù)采集成為常態(tài)，部分學(xué)習(xí)者獲得幫助其學(xué)習(xí)所需信息的機(jī)會(huì)有限，并且無(wú)法生成和共享電子數(shù)據(jù)，在建立個(gè)性化學(xué)習(xí)途徑以確定和滿足個(gè)人的需求時(shí)，可能處于不利地位。

4.2 倫理方面

目前，私營(yíng)公司在將人工智能的力量引入教育領(lǐng)域方面起著帶頭作用，這引起了人們對(duì)學(xué)生數(shù)據(jù)的隱私、保護(hù)和使用的擔(dān)憂。對(duì)人工智能的倫理基礎(chǔ)的考慮不僅僅是在人工智能服務(wù)和解決方案的部署階段，而是應(yīng)該貫穿在整個(gè)開發(fā)過程中。我們?nèi)绾未_保在編制方案時(shí)不引入性別、種族、社會(huì)經(jīng)濟(jì)和能力偏見？我們?nèi)绾未_保社會(huì)和文化的陳詞濫調(diào)不被傳播？我們?nèi)绾未_保所有學(xué)習(xí)者無(wú)論身處何處，都能同樣享受到人工智能帶來(lái)的好處？

4.3 技術(shù)依賴

增加對(duì)人工智能的依賴并不完全有利于我們。我們可以預(yù)期，在發(fā)達(dá)國(guó)家和發(fā)展中國(guó)家，自動(dòng)化將取代越來(lái)越多的工作崗位。還有一種危險(xiǎn)，即將某些任務(wù)和認(rèn)知功能委托給機(jī)器可能會(huì)增加我們對(duì)技術(shù)的依賴，同時(shí)削弱人類執(zhí)行這些任務(wù)的能力，也許這就是“要么使用，要么失去”。依賴電腦鍵盤的學(xué)生會(huì)失去清晰書寫的能力嗎？那些使用電子表格或計(jì)算器進(jìn)行計(jì)算的人在心算方面是否變得不那么熟練？對(duì)GPS的依賴是否影響了我們的導(dǎo)航能力？隨著這些技術(shù)變得越來(lái)越智能、越來(lái)越成熟，了解我們想要去哪里也將變得比以往任何時(shí)候都重要。

4.4 持續(xù)的專業(yè)發(fā)展

為了使人工智能有益于教育的發(fā)展，有必要讓教育者和管理者掌握評(píng)估和解釋數(shù)據(jù)結(jié)果的技能。

有效的人工智能解決方案產(chǎn)生易于理解的可視化表和儀表盤，讓用戶可以實(shí)時(shí)查詢數(shù)據(jù)并吸收見解，可以很容易地生成針對(duì)不同角色和功能定制自定義視圖和報(bào)表。在此背景下，持續(xù)的專業(yè)發(fā)展的核心重點(diǎn)是培訓(xùn)教育者學(xué)會(huì)將人工智能所產(chǎn)生的見解融入教學(xué)實(shí)踐，并幫助他們?yōu)閷W(xué)生創(chuàng)建個(gè)性化的學(xué)習(xí)途徑。

5. 未來(lái)展望

5.1 做事方式變革

5.1.1 元認(rèn)知障礙

人工智能可以在元認(rèn)知“腳手架”中發(fā)揮作用，讓學(xué)生自己從學(xué)習(xí)過程產(chǎn)生的洞察力中獲益。學(xué)生將越來(lái)越成為人工智能解決方案和服務(wù)的主要用戶，而不僅僅是教育者、管理員和系統(tǒng)所有者分析數(shù)據(jù)的對(duì)象。學(xué)習(xí)者對(duì)自己的學(xué)習(xí)方式和思維方式有更深入的了解，就會(huì)對(duì)自己的教育有更大的能動(dòng)性和控制力，并且具備重要的自我知識(shí)也是終身學(xué)習(xí)的基石。

5.1.2 個(gè)性化評(píng)估和認(rèn)證

人工智能還將在應(yīng)對(duì)教育技術(shù)專家面臨的下一個(gè)重大挑戰(zhàn)上發(fā)揮重要作用，即如何支持個(gè)性化評(píng)估和認(rèn)證。我們知道，目前高校的評(píng)估形式很少與學(xué)生進(jìn)入職場(chǎng)時(shí)所需的技能相一致。21世紀(jì)環(huán)境中所需的一些高階思維技能，如記憶、比較、分析和推理，可以用傳統(tǒng)的標(biāo)準(zhǔn)化考試來(lái)評(píng)估，但軟技能、人際交往技能、道德品質(zhì)、團(tuán)隊(duì)合作、協(xié)作等很難用這種形式來(lái)評(píng)估。

越來(lái)越多的機(jī)構(gòu)和學(xué)校正在研究使用微型證書來(lái)衡量復(fù)雜和一般的學(xué)習(xí)能力，包括21世紀(jì)技能中的批判性思維、創(chuàng)造力、溝通技能和創(chuàng)業(yè)精神等技能。如果這些新證書在引入后仍具有相關(guān)性，人工智能可以對(duì)其效力和有效性進(jìn)行持續(xù)分析，并針對(duì)利益相關(guān)者的興趣或?qū)I(yè)領(lǐng)域定制實(shí)時(shí)儀表盤展示這些發(fā)現(xiàn)和見解。