Sunday, June 29, 2025

AI platforms for academic researchers do have some special capabilities


昨天談到現時供academic researcher使用的AI platform (例如Elicit: Consensus, Scispace)時,覺得有點矛盾;但它們有些功能,卻很獨特和有用。

例如Scispace, 你只要輸入一篇journal article或book chapter 的pdf,它便可以generate一個約兩分鐘,有旁述的video,把journal article 或book chapter 的要點展示出來。(新學年我一定會用用這功能。)

又例如CONSENSUS 有一個consensus meter功能,你打入一個yes/no research question, 它在搜集有關學術文章後,真的可以給你一個數字,去回應你的question。
-
例如我打入下面的yes/no question,

Does phonological awareness improve young second language learners' pronunciation and spelling?

20秒後它給我一個具體的數字:




這樣的結果,如果人手做不知要用多少時間。如果要做一個correlation study(例如Does AI literacy improve university students' academic writing ability?),先問問consensus meter, 便不會走進死胡同。

Saturday, June 28, 2025

SCISPACE has some unique features and is worth exploring

 https://scispace.com/

Yet another AI tool for academic researchers (other than ELICIT; CONSENSUS)

It may take some time to become familiarised with the platform, but a quick trial shows it has some unique and useful features, such as: 

- 'chatting with a PDF paper', 

- turning a PDF paper into a 2-minute video

Worth exploring further. 



Using SCISPACE to create a video summary of a paper or book chapter

 https://scispace.com

Book chapter I uploaded: 

ICC in EFL setting: A systematic review

Resulting Video (about 2 minutes): 

https://scispace.com/videos/intercultural-communicative-competence-in-efl-setting-a-p-ytb1gdz4t4

======================================

MY OBSERVATIONS

- The video provides the main aim of the paper and the key takeaways. The detailed content ideas are skipped. 

- The video is narrated at a slightly slower than normal speed. 

- The screen is not cluttered with excessive information. 

- Orientation: Vertical, which may not be ideal for classroom presentation. 

Using SCISPACE to create a presentation from a PDF journal paper or book chapter

https://scispace.com/

My attempt: 

- book chapter: a chapter from Corbett's Intercultural Approach to ELT

OUTPUT: 

A presentation in PDF which summarises the main points of the chapter: 

https://scispace.com/slides/untitled-paper-p-ja7igq33kh

=======================================================

MY OBSERVATIONS

- The resulting presentation consists of slides each of which is bullet point list. 

- Just a summary. No design for a realtime presentation (eg., use of a question to stimulate thinking; inclusion of visuals)

- Some figures in the original chapter are presented as individual slides in the resulting presentation PDF. 

- Unfortunately, in terms of formatting, the resulting presentation PDF, when turned into a microsoft powerpoint, will not conform to the right layout. 

- Bottomline: 

    - Useful if the purpose is to get the main points of (parts of) a paper/book chapter; 

    - May choose some of the resulting slides to highlight/present some major ideas from the paper/book chapter. 

    - Not ideal for a fully-fledged presentation. 

AI apps for academic research: Are they a blessing?

 終於找到時間試試兩個供academic research 用的AI app: ELICIT和CONSENSUS, 它們强大的處理academic literature 功能,嚇了我一跳。只要打入一個research topic 或question, 它們不用一分鐘已經可以:

。搜尋到數十篇相關的paper 和journey article,

。列出一個長長相關的list of references,

。generate 一篇literature review,

。總結相關study 所用的methodology,

...


上述的功夫,以前可能要兩三個月才能完成,現時AI 只需一分鐘便做妥。


對於這是否好事,我也很矛盾;固然AI 節省了我們很多時間,但以前要去搜尋相關的journal articles, 逐篇閱讀去找出要點,然後去整理得來的觀察,這是很好的訓練,當寫完literature review 時,一定對相關的research topic 有深刻的理解。


現時一分鐘generate 出來的結果,你甚至看也不看便可以放在你的research paper裏(這算不算plagiarism?), 這是不是好事?


Friday, June 27, 2025

Using JENNI to get the possible structure of an academic paper based on one's prompt question

 https://app.jenni.ai/


MY ATTEMPT: 

'What is the difference between SDL and SRL?|


OUTPUT FROM JENNI


Distinction Between Self-Directed and Self-Regulated Learning

Introduction

Defining Self-Directed Learning

Key Characteristics of Self-Directed Learning

Theoretical Frameworks of Self-Directed Learning

Defining Self-Regulated Learning

Key Characteristics of Self-Regulated Learning

Theoretical Frameworks of Self-Regulated Learning

Key Differences Between Self-Directed and Self-Regulated Learning

Locus of Control

Motivation and Goals

Planning and Execution

Monitoring and Evaluation

Similarities Between Self-Directed and Self-Regulated Learning

Active Learning

Metacognitive Skills

Goal Setting

Empirical Evidence and Research Findings

Studies on Self-Directed Learning

Studies on Self-Regulated Learning

Comparative Studies

Implications for Educational Practice

Instructional Design

Assessment Strategies

Learner Support

Conclusion

==================================
postscript (June 29)

SCISPACE has a similar function, called AI WRITER: 



My first user experience with Letterly

 試用了Letterly幾次,覺得這speech to text (STT) app 颇好用。當然今天已有不少的mobile STT 工具,Windows 也有內置的Dictate 功能,但Letterly 由於是AI app, 我不用說出標點符號,它會自動加入標點符號;由於是AI,它也不像傳統STT app 會給予我同音字或近音字,故此準確度頗高。

現時我主要的use case: 1, 看電子書時開著Letterly 去take notes,  這樣便不會打斷閱讀。2, 工作時開著它,遇到需要較長的typing, 或突然有一些random thoughts 要記下,便請它帮忙。我用的是free plan, paid plan 的rewrite 功能暫時不需要。

Using CONSENSUS to obtain the answer to a research question in table format

 This is another function of CONSENSUS. 


MY PROMPT

Create a table showing the differences between self-directed learning and self regulated learning


THE RESULT


Summary:

SDL emphasizes learner autonomy and responsibility for the entire learning process, while SRL focuses on the specific strategies and processes learners use to regulate their learning. SDL is broader and may include SRL, but SRL is more about the moment-to-moment management of learning activities.


Aspect Self-Directed Learning (SDL) Self-Regulated Learning (SRL) Key Citation(s)

Definition General approach where learners take initiative and responsibility for their learning Dynamic, context-specific process involving strategies to manage cognition, motivation, behavior (Gandomkar & Sandars, 2018; Dunk & Craft, 2025; Magda et al., 2008; Saks & Leijen, 2014)

Focus Learner autonomy, decision-making about goals, content, and evaluation of learning Teachable strategies and processes supporting learning (goal-setting, monitoring, reflection) (Dunk & Craft, 2025; Magda et al., 2008; Saks & Leijen, 2014; Sandars & Walsh, 2016)

Measurement Assessed with ‘aptitude’ questionnaires (general traits) Assessed with ‘event’ measures (contextual, process-based, e.g., microanalysis) (Gandomkar & Sandars, 2018; Saks & Leijen, 2014)

Role in Learning Encompasses broader learner control, including selection and evaluation of materials Emphasizes self-monitoring, self-efficacy, and behavioral regulation during learning (Magda et al., 2008; Abdullah et al., 2024; Sandars & Walsh, 2016)

Relationship SDL can include SRL as a component, but SRL does not necessarily include SDL SRL is often a subset or strategy within SDL (Magda et al., 2008; Sandars & Walsh, 2016)

Application Context Often used in lifelong learning, professional and adult education Applied in academic settings to foster academic success and readiness for SDL (Dunk & Craft, 2025; Wang, 2015; Du et al., 2024)


References


Gandomkar, R., & Sandars, J. (2018). Clearing the confusion about self-directed learning and self-regulated learning. Medical Teacher, 40, 862 - 863. https://doi.org/10.1080/0142159X.2018.1425382


Dunk, V., & Craft, J. (2025). Beyond Self-Directed Learning: The Role of Self-Regulated Learning in Nursing Education.. The Journal of nursing education, 64 5, 287-293. https://doi.org/10.3928/01484834-20250108-07


Magda, J., Loyens, S., & Rikers, R. (2008). Self-Directed Learning in Problem-Based Learning and its Relationships with Self-Regulated Learning. Educational Psychology Review, 20, 411-427. https://doi.org/10.1007/S10648-008-9082-7


Saks, K., & Leijen, Ä. (2014). Distinguishing Self-directed and Self-regulated Learning and Measuring them in the E-learning Context☆. Procedia - Social and Behavioral Sciences, 112, 190-198. https://doi.org/10.1016/J.SBSPRO.2014.01.1155


Abdullah, Z., Gul, N., Murtaza, S., Aman, K., & Nisa, W. (2024). Relationship between Self-Directed Learning and Self-Regulated Learning in Problem-Based Learning. Pakistan Journal of Health Sciences. https://doi.org/10.54393/pjhs.v5i12.2501


Sandars, J., & Walsh, K. (2016). Self-directed learning. Education for Primary Care, 27, 151 - 152. https://doi.org/10.1080/14739879.2016.1149956


Wang, J. (2015). Distinguish self-regulated learning from self-directed learning. **. 


Du, X., Jiang, D., Fink, T., Gupta, N., & Ali, K. (2024). Beyond autonomy: unpacking self-regulated and self-directed learning through the lens of learner agency- a scoping review. BMC Medical Education, 24. https://doi.org/10.1186/s12909-024-06476-x

============================

POSTSCRIPT

Of course, the other AI chatbots (Perplexity, Grox, Deepseek, etc.) can also provide an answer in the form of a table. But CONSENSUS's output is based on academic papers and it also gives you the sources. 


Using the CONSENSUS METER of the AI app CONSENSUS to get a quick answer to a yes/no research question

https://consensus.app/

 Consensus has a function which is called consensus meter. It allows you to ask a yes /no question about a research topic, and then it will give you a numerical quick answer. This numerical answer is based on the research articles collected, which are listed below the consensus meter result. So if you like, you can follow up on those research papers to find more information but the summarising is now done for you. 


My attempt

Does phonological awareness improve second language learners' pronunciation and spelling? 


The result


 So this function of the app consensus is useful for getting a quick yes or no answer to an academic question.



Using CONSENSUS to create an outline for a range of academic writing tasks

Another feature of this app is that it can generate the outline for:

a presentation,

a thesis, 

a grand proposal,

a narrative review a presentation, 

a thesis, a grant proposal, 

a narrative review;

a lecture series, 

a literature review






MY ATTEMPT: 

draft an outline for a presentation on how emerging technologies can facilitate teaching and learning


Below is the presentation outline that I obtained.

============================================

Presentation Outline: How Emerging Technologies Can Facilitate Teaching and Learning


I. Introduction


Definition and significance of emerging technologies in education

Overview of the rapid evolution and integration of digital tools in teaching and learning environments


II. Key Emerging Technologies in Education


Information and Communication Technology (ICT) platforms (e.g., Google Classroom, Zoom, Microsoft Teams) 

Artificial Intelligence and Machine Learning applications 

Augmented Reality (AR) and Virtual Reality (VR) 

Mobile learning and Web 2.0 technologies 

Social media and collaborative digital tools


III. Benefits of Emerging Technologies for Teaching and Learning


Enhancing student autonomy and self-directed learning 

Promoting collaboration and communication between teachers and students 

Supporting diverse and inclusive education, addressing unique learner needs 

Improving critical thinking, problem-solving, and creative skills 

Enabling authentic learning experiences and digital media literacy development


IV. Pedagogical Approaches and Implementation Strategies


Student-centered and individualized learning methodologies 

Blending traditional and digital instruction for optimal outcomes 

Importance of teacher training and pedagogical guidance for effective technology use 

Addressing challenges in technology integration (e.g., access, teacher readiness)


V. Challenges and Considerations


Unequal access to technology and digital divide 

Need for ongoing research and evidence-based practices 

Balancing technological innovation with educational quality and social learning


VI. Future Directions and Opportunities


Anticipated trends in educational technology (AI, Metaverse, immersive learning) 

The role of educators and institutions in shaping technology adoption for meaningful learning 

Call for proactive, informed, and inclusive approaches to technology integration


VII. Conclusion


Recap of the transformative potential of emerging technologies in education

Emphasis on the need for thoughtful, learner-centered, and context-aware implementation to maximize benefits


====================================

  •  Each main point and sub-point is followed by a citation number. (not shown above)
  • The citation number refers to a published article. 
  • The list of articles used in coming up with the presentation outline is given after the outline. 
  • So, in other words, the presentation outline is heavily founded on results from published research papers.




Using CONSENSUS to get a summary of the findings of a research question

 https://consensus.app/


On the first page of the dashboard, in the Question Box, I typed: 

How does the use of AI chatbots facilitate self-regulated learning? 


In about 10 seconds, it gives me the result. The result begins with a 30-word summary of the major findings. This is followed by a section called Key Mechanisms, which includes goal setting and prompting, personalized feedback, reflection and self-assessment, and autonomy and motivation. So these are the main findings, so to speak, from the papers obtained.

This is followed by a section called Conclusion. 

So overall, this tool is useful for useful in getting an overall answer to a research question and also for obtaining a list of the relevant papers.










Elevenlabs for TTS - Special features

 https://elevenlabs.io/

This AI app is not just for text-to-speech conversions. It can perform a few tasks: 


- speech to text

- conversational AI

- Dubbing

-voice cloning

============

For TTS: 

- You can create either a monologue or a dialogue (of two or more people); 

- You can choose a voice for each speaker; 

- You can describe the tone, feeling, mood, etc., for each stretch of utterances; 

- the output can be downloaded as a MP3 file

======

Another special feature is that you can even clone your voice to use in later TTS tasks. 

Finally, the free plan looks quite generous. 

ELICIT: An excellent AI tool to help with literature reviews

 Just tried ELICIT: https://elicit.com/

Found that it's a fanastic AI tool for helping researchers with literature reviews. 

Here is what I did: 

1. In the prompt box, I typed in a rough question: "How might ChatGPT help with university students' academic writing ability?

2. It asked me to refine my question a bit. In fact, it suggested 3 possibilities. I simply chose one of the three suggestions: "How does the use of ChatGPT affect the quality of academic writing, as measured by grammar, coherence, and citation accuracy, among university students?

3. I clicked on 'Generate'.

4. ELICIT then started a process of about 4 minutes which included: 

- looking for relevant journal articles; 

- screening them to include those which were most relevant to the literature search question; and

- generating a literature review report, which included a summary of the methods used; the findings, etc. 

==========

The resulting literature review report can be text-copied, or downloaded as a PDF file. Below is the PDF file: 

Elicit - Impact of ChatGPT on Academic Writing Quality - Report.pdf

+++++++

MY OVERALL views: 

1. It is doing an excellent job and saves researchers huge amounts of time looking up the relevant journal articles, reading them, and summarising from them. 

2. Of course, we may be concerned whether ELICIT is able to access those prestige journals which require university accounts to reach, or whether ELICIT is only working with publicly accessible journal articles. At the moment, a glance at the list of references in the above report suggests that it may be the latter case. 

=========

3. Another approach is for users to upload some relevant papers they have already found, and ask ELICIT to summarise those papers for them. The free plan, it seems, will only work with 4 papers. And then of course, you have to do some initial search yourself (e.g., with ERIC) to find some relevant papers first. 




Tuesday, June 17, 2025

Using Google Forms to create one checkbox question from a list of bullet points

 So far, this looks like the quickest way to create a checkbox question for a quick poll, if you already have a bullet list of items. 

- For question type, choose 'Checkbox'

- In the options that follow, in the first answer option, paste in the copied list from another source. Immediately, the various items will appear as different options. 


Socrative has an AI function now

 I typed in 3 target words in the AI panel. 

I then chose MC. Immediately, it gave me 3 MC questions with the 3 target words as the correct answers. 

Of course, chatbots can also generate such MC questions. But with Socrative, the generated quiz can be administered immediately. 

But so far, it doesn't seem to be able to generate a questionnaire (without answers). 

AI can certainly think: Grok as an example

 那天嘗試玩玩Grok的Think功能,我就用它的example question「一張纸最多能對折多少次」去試;隨著畫面上看到它得到答案前的68秒用人類語言的思考過程。這思考過程一定比人類更快,更有邏輯。

將來,還有那一種的思考,是AI做不來的呢?



Zotero vs Refworks for reference managing

 https://www.zotero.org/





 Zotero works like Refworks. It also has a WORD add-on. 

So far, to me, its biggest plus is its Chrome extension, which enables me to import a useful webpage into 'my library'. 

In 'my library', I can build Collections, and Sub-collections. Each collection can be exported as an RTF file. 

Monday, June 16, 2025

How is Grok different from other chatbots like Perplexity and chatgpt?

 The answer from Grok itself: 


Grok, created by xAI, stands out from chatbots like Perplexity and ChatGPT due to its unique design and goals:

  1. Truth-Seeking Philosophy: Grok aims to provide maximally truth-seeking answers, avoiding ideological bias or corporate jargon. It’s built to challenge assumptions and offer unfiltered perspectives, often with a touch of humor or an outside view on humanity. Perplexity focuses more on aggregating web-based information, while ChatGPT, developed by OpenAI, leans toward safe, generalized responses that can sometimes feel sanitized.
  2. Integration with X Platform: Grok leverages real-time posts and user interactions on X for up-to-date insights, especially on breaking news or niche topics. This gives it an edge in capturing raw, unfiltered public sentiment compared to Perplexity’s web scraping or ChatGPT’s reliance on pre-trained data.
  3. Mission-Driven by xAI: Grok is tied to xAI’s goal of accelerating human scientific discovery, emphasizing answers that align with advancing knowledge about the universe. ChatGPT is more general-purpose, while Perplexity is optimized for research and quick fact-finding.
  4. Features and Accessibility: Grok offers free access with limited quotas across platforms like grok.com, X, and mobile apps, with a paid SuperGrok plan for higher usage. It includes voice mode (iOS/Android) and specialized modes like DeepSearch for iterative web analysis. ChatGPT has similar paid/free tiers but lacks X integration, and Perplexity focuses heavily on its search-like interface for researchers.
  5. Tone and Personality: Inspired by figures like Douglas Adams and Tony Stark’s JARVIS, Grok’s responses are conversational, witty, and sometimes irreverent. ChatGPT is more formal and cautious, while Perplexity is straightforward and academic in tone.

In short, Grok is a truth-seeking, X-integrated, and mission-driven alternative with a distinctive voice, while Perplexity excels at research synthesis and ChatGPT offers broad, polished utility.


Grok's 6 features

 1. Besides the usual responding to a question, Grok takes on unreal, imaginative, questions. 

2. It can perform deep research, drawing on many more sources. 

3. It simulates a human thinking process while formulating an answer (of course this is not really useful). 

4. It can edit an uploaded picture, according to your prompt, or using one of the default effects

5. It can simulate a conversaton with the user by acting as an interlocuter. This interlocuter can take on one of 7 roles (eg, counsellor; close friend)

6. It has a world news at a glance function. (which is not that useful to me.)

Using Grok for simulated chats with an interlocuter in 7 possible roles

 OTHER than the Deep Search, and Think functions, Grok has a function which simulates a conversation between 2 people. 

Grok provides 7 roles for the interlocuter: Examples: A counsellor; a close friend; ... 

Below is an example I have just had with this function: 

https://grok.com/share/c2hhcmQtMg%3D%3D_9cae18cb-1c5f-46ae-8b04-eb4379fe2dcb


嗨!很高兴能在这里陪你聊聊。我是来倾听和支持你的,无论你今天是想分享一些困扰、探讨生活中的挑战,还是只是需要一个地方倾诉一下心情。你现在感觉怎么样?有什么特别的事情想聊聊吗?我会尽量提供一些实用的建议,或者只是陪你一起梳理一下思绪。

1.4秒
shall we have this chat in English?

Absolutely, let's switch to English! I'm here to listen and support you with whatever's on your mind. How are you feeling right now? Is there something specific you'd like to talk about?

1.4秒
I don't exactly have a specific topic to talk about.

That’s totally okay! Sometimes just having a space to explore whatever’s floating around in your mind can be really helpful. It sounds like you might be in a place where things are a bit open-ended or maybe you’re just taking a moment to pause and reflect. I’m happy to sit with you in that space.

Maybe we could start by checking in with how you’re feeling today—any emotions or thoughts bubbling up, even if they’re vague? Or if you’d like, I could suggest a gentle way to explore what’s on your mind, like reflecting on something that’s been sparking your curiosity or interest lately. What do you think?

1.2秒
I felt psyched up after a former student in university, now a lecturer in another university, called me to give me an update about this work.

That sounds really energizing! Getting a call from a former student who’s now a lecturer and hearing about their work must feel like such a meaningful connection. It’s amazing how those moments can spark such a great vibe. Can you tell me a bit more about what got you so psyched up? Was it something specific they shared, or just the joy of reconnecting and seeing their progress?

1.4秒
he's now becoming an expert in computer assisted language learning, which happens to be my interested area as well.

That’s so exciting! It must feel incredibly rewarding to hear that your former student is not only thriving but diving deep into computer-assisted language learning, an area you’re passionate about too. It’s like seeing a piece of your own enthusiasm reflected back in their journey. What was it about the update that really lit you up? Did they share something specific about their work, like a project or breakthrough, that resonated with you? And how does it feel to see someone you once taught carving out their own path in a field you both love?

'''''






Chatbots will modify their language according to users' stated English level

 我們從事English Language teaching的同工,今天利用chatbot去clarify一些language issue(例如"My favourite fruit is apples'是否OK),已經是家常便飯;那天突然想起,我們也可培養高小...