عنوان مقاله:
Rethinking the entwinement between artificial intelligence and human learning: What capabilities do learners need for a world with AI?
بازنگری در آمیختگی بین هوش مصنوعی و یادگیری انسان: یادگیرندگان برای دنیایی با هوش مصنوعی به چه قابلیت هایی نیاز دارند؟
رشته: مهندسی کامپیوتر
گرایش: هوش مصنوعی
دانلود رایگان این مقاله:
دانلود مقاله هوش مصنوعی و یادگیری انسان
مشاهده سایر مقالات جدید:
2. Defining the territory
efore starting our polylogue, we need to clarify our terminology. Different terms have been used in the literature to describe what people need to know and be capable of doing to function successfully in society, such as ‘literacy’, ’skills’, ’competencies’ and ’capabilities’. The term ‘literacy’ historically has been associated with one’s ability to read and write. The initial technical notions of literacy as the ability to use the alphabet have been replaced with the functional notions of literacy as the ability to use technical skills to pursue personal goals and function within society. For example, the recent OECD (2019) report on adult skills describes literacy as “the ability to understand, evaluate, use and engage with written texts to participate in society, to achieve one’s goals, and to develop one’s knowledge and potential” (p. 18). Similar notions of literacy have been applied to conceptualise various technology-related abilities, such as ‘ICT literacy’ and ‘digital literacy’ (see for review Markauskaite, 2006). A similar view of literacy is used in the context of AI (Long & Magerko, 2020). The term ‘skills’ rather than ‘literacy’ has become more common recently, particularly in discussions about ‘the 21st-century skills’ or ‘generic skills’ and in professional education and lifelong learning contexts (OECD, 2019). This term, however, has been heavily criticised in educational literature. This critique centred on two main aspects. First, the term ’skills’ is usually associated with systematic instruction and pre-specified measurable levels of achievement. Thus, it is often too specific to address the unpredictability of what people will need to be capable of doing in the future. Secondly, when this term is used to refer to ‘future proof’ or ‘21st-century skills’, it usually includes traits or personal characteristics (e.g., creativity) rather than skills (Kirschner & Stoyanov, 2020), and is thus viewed as being semantically inaccurate. Such literature usually proposes a broader term ‘competency’ as a more appropriate term in future-oriented contexts (cf. Buckingham Shum & Deakin Crick, 2016). Other literature, however, assigns little importance to the differences between these terms. For example, the OECD (2019) report on adult skills mentioned above uses the terms ‘skills’ and ‘competency’ synonymously even if it acknowledges that ‘competency’ is a broader term that includes “knowledge, skills and attitudes (beliefs, dispositions, values)" and refers to “the application and use of knowledge and skills in common life situations as opposed to the mastery of a body of knowledge or a repertoire of techniques” (p. 98). In more recent times, broader terms such as ‘capacity’ and ‘capability’ have been commonly used (Gangas, 2016; Markauskaite & Goodyear, 2017; Poquet & de Laat, 2021). These terms primarily refer to the human qualities and potential to do certain things and achieve desired outcomes. It shifts the focus from the demonstrated behaviours to the potential, dispositions and opportunities within one’s reach to pursue specific values and outcomes. For inclusiveness, we adopted the broader term ’capabilities’ rather than ‘literacy’, ‘skills’ or ‘competency’ in this paper.
This is in line with our aim to explore the space of what people should be capable of doing to succeed in a world with AI, rather than to provide one specific definition of what capabilities directly related to AI entail. Although it is certainly important to understand the latter capabilities as well, this work has partly been done by others. For example, Long and Magerko (2020), drawing on an extensive scoping literature review, explored the notion of AI literacy. They defined ‘AI literacy’ as “a set of competencies that enables individuals to critically evaluate AI technologies; communicate and collaborate effectively with AI; and use AI as a tool online, at home, and in the workplace” (p. 598). Their review resulted in 17 core competencies related to people’s understanding of what AI is, what AI can do, how AI works, how AI should be used, and how people perceive AI. However, the issues induced by AI reach far beyond skills and knowledge or attitudes directly related to AI, to include characteristics and competencies that have been critical for many previous generations but now take on new shapes, such as cooperation, creativity, complex problem-solving, flexibility and change (Buchanan et al., 2018; Markauskaite, 2020). In short, an AI-centred view of capabilities may not capture many other capabilities that learners need to develop for a world with AI. This inevitably brings the danger of being too broad and answering the question of “What kinds of capabilities are needed for a world with AI” by saying “The same ones as always”. Even so, it is important to consider how these capabilities change in an AI context. What is distinctive about AI-based technologies is the ability to automate certain processes and emulate (even exceed in some cases) human performance. It is essential to consider new possibilities and barriers one might encounter in enacting and enhancing those capabilities that now become distributed between humans and intelligent machines when humans and machines perform in cooperation. In this context, we asked the experts to address the above questions about capabilities that people need in a world with AI.
3. Methodology
In this multi-authored paper, we adopted a polychronic and polyphonic research approach, similar to those used for collective knowledge-making in experimental postdigital dialogues (Jandri´c et al., 2019; Matusov, Marjanovic-Shane, & Gradovski, 2019). Jandri´c et al. (2019), drawing on Peters (2015) work, describe postdigital dialogues as a form of ‘collective intelligence’: “a scientific, technical and political project that aims to make people smarter with computers, instead of trying to make computers smarter than people. So, collective intelligence is neither the opposite of collective stupidity nor the opposite of individual intelligence. It is the opposite of artificial intelligence. It is a way to grow a renewed human/cultural cognitive system by exploiting our increasing computing power and our ubiquitous memory.” (Peters, 2015, p. 261 cited in; Jandri´c et al., 2019, p. 164). L. Markauskaite et al. Computers and Education: Artificial Intelligence 3 (2022) 100056 3 One may question if the opposition of human and artificial intelligence is necessary. However, given the complexity of the challenge that we set out to explore, and the importance of multiple human perspectives in finding acceptable solutions, this approach was appropriate. Over about three months, we engaged in an orchestrated written conversation, characterised by a polychronic organisation of our writing, which occurred non-linearly in the form of asynchronous dialogue. The collaborative writing was guided by the polyphonic structure, aiming to ensure that an independent voice of each author representing a particular intellectual tradition is initially heard, and then juxtaposed with other voices and attuned to each other. Before starting this polylogue, we worked together as a part of a larger multidisciplinary team for about 12 months on creating intellectual foundations for a joint project, “Empowering learners in the age of AI”. The first author, who orchestrated the dialogue, invited team members with expertise in different domains to participate in a jointly written polylogue to discuss capabilities that students need for a world with AI. These members, who became co-authors, had backgrounds in diverse disciplinary fields (e.g., education, learning sciences, computer science, and engineering) and represented different conceptual perspectives towards the capabilities and AI in education. They were chosen seeking to ensure representation of different career stages and genders. Each co-author was able to invite their collaborators representing a similar conceptual perspective to participate in the polylogue alongside. One additional co-author joined the team at this stage. The polylogue took place online using Google Docs for collaborative writing. It spanned three phases. In Phase 1, each co-author was asked to adopt a perspective representing their domain of expertise and respond to a set of five questions. In their responses, they were asked to articulate their perspective and describe: 1) what kind of capabilities will people need in a world with Al, 2) how these capabilities could be conceptualised, 3) how they could be developed, 4) how this development could be empirically studied and assessed, and 5) what else should be considered when we think about how to prepare people for a world with AI. To ensure that all voices are heard and avoid ‘groupthink’, each co-author was asked to write their initial contribution independently and not read the contributions of other co-authors before drafting their responses. At the end of this phase, the first coordinating author integrated all responses to each question and made some editorial comments asking authors to clarify their ideas when necessary. She also identified the initial overarching themes, including the dominant orientation of each perspective, and drafted openers for the joint discussion. In Phase 2, all authors were invited to 1) read each other’s contributions and leave any questions and comments for their co-authors; 2) read peer comments and make changes that they deem necessary in their answers, and 3) reflect on everyone’s contributions and add their insights to the joint discussion. All authors were invited to co-write this section by integrating their ideas and critique while respecting each other’s points. The first author lightly edited the jointly produced text and submitted it for peer review.
(دقت کنید که این بخش از متن، با استفاده از گوگل ترنسلیت ترجمه شده و توسط مترجمین سایت ای ترجمه، ترجمه نشده است و صرفا جهت آشنایی شما با متن میباشد.)
2. تعریف قلمرو
قبل از شروع چند گفتاری خود، باید اصطلاحات خود را روشن کنیم. اصطلاحات مختلفی در ادبیات برای توصیف آنچه افراد باید بدانند و قادر به انجام آنها برای عملکرد موفقیت آمیز در جامعه هستند، مانند "سواد"، "مهارت ها"، "شایستگی ها" و "قابلیت ها" استفاده شده است. اصطلاح «سواد» از نظر تاریخی با توانایی فرد در خواندن و نوشتن همراه بوده است. مفاهیم فنی اولیه سواد به عنوان توانایی استفاده از الفبا با مفاهیم کاربردی سواد به عنوان توانایی استفاده از مهارت های فنی برای پیگیری اهداف شخصی و عملکرد در جامعه جایگزین شده است. به عنوان مثال، گزارش اخیر OECD (2019) در مورد مهارت های بزرگسالان، سواد را به عنوان "توانایی درک، ارزیابی، استفاده و تعامل با متون مکتوب برای مشارکت در جامعه، دستیابی به اهداف، و توسعه دانش و پتانسیل خود" توصیف می کند. 18). مفاهیم مشابهی از سواد برای مفهوم‌سازی توانایی‌های مختلف مرتبط با فناوری، مانند «سواد فناوری اطلاعات و ارتباطات» و «سواد دیجیتال» به کار رفته است (برای بررسی Markauskaite، 2006 مراجعه کنید). دیدگاه مشابهی از سواد در زمینه هوش مصنوعی استفاده می شود (Long & Magerko, 2020). اصطلاح «مهارت‌ها» به جای «سواد» اخیراً رایج‌تر شده است، به‌ویژه در بحث‌هایی درباره «مهارت‌های قرن بیست و یکم» یا «مهارت‌های عمومی» و در زمینه‌های آموزش حرفه‌ای و یادگیری مادام‌العمر (OECD، 2019). این اصطلاح اما در ادبیات آموزشی به شدت مورد انتقاد قرار گرفته است. این نقد بر دو جنبه اصلی متمرکز بود. اول، اصطلاح «مهارت‌ها» معمولاً با آموزش منظم و سطوح پیشرفت قابل اندازه‌گیری از پیش تعیین شده مرتبط است. بنابراین، پرداختن به غیرقابل پیش‌بینی بودن آنچه که مردم در آینده باید انجام دهند، بسیار خاص است. ثانیاً، هنگامی که این اصطلاح برای اشاره به «اثبات آینده» یا «مهارت‌های قرن بیست و یکم» استفاده می‌شود، معمولاً شامل ویژگی‌ها یا ویژگی‌های شخصی (مثلاً خلاقیت) به جای مهارت‌ها می‌شود (Kirschner & Stoyanov, 2020) و بنابراین به عنوان از نظر معنایی نادرست بودن چنین ادبیاتی معمولاً یک اصطلاح گسترده‌تر «شایستگی» را به‌عنوان یک اصطلاح مناسب‌تر در زمینه‌های آینده‌محور پیشنهاد می‌کند (ر.ک. Buckingham Shum & Deakin Crick، 2016). با این حال، ادبیات دیگر اهمیت کمی به تفاوت بین این اصطلاحات می دهد. برای مثال، گزارش OECD (2019) در مورد مهارت‌های بزرگسالان که در بالا ذکر شد، از اصطلاحات «مهارت‌ها» و «شایستگی» مترادف استفاده می‌کند، حتی اگر تصدیق کند که «شایستگی» یک اصطلاح گسترده‌تر است که شامل «دانش، مهارت‌ها و نگرش‌ها (باورها، تمایلات، ارزش‌ها)" و به "کاربرد و استفاده از دانش و مهارت‌ها در موقعیت‌های معمول زندگی بر خلاف تسلط بر مجموعه‌ای از دانش یا مجموعه‌ای از تکنیک‌ها" اشاره دارد (ص 98). در زمان‌های اخیر، اصطلاحات گسترده‌تری مانند «ظرفیت» و «قابلیت» معمولاً مورد استفاده قرار گرفته‌اند (Gangas, 2016; Markauskaite & Goodyear, 2017; Poquet & de Laat, 2021). تمرکز را از رفتارهای نشان‌داده‌شده به پتانسیل، استعدادها و فرصت‌های در دسترس فرد برای دنبال کردن ارزش‌ها و نتایج خاص تغییر می‌دهد. برای فراگیری، ما از اصطلاح گسترده‌تر «قابلیت‌ها» به جای «سواد»، «مهارت‌ها» یا «شایستگی» در این زمینه استفاده کردیم. کاغذ است
