https://glow.digital.moe.gov.sg/collection/019d4329-9e62-7848-ad34-cd2989310d45
for my own learning.
1 One
Based on the page you are viewing, here is a summary of the mandatory learning bite from the Ministry of Education (MOE):
Overview: AI Ep 1: Why Should You Care About AI?
This is a mandatory module for MOE educators and staff created by the MOE Digital Excellence & Products Division (DXD).
Deadline: 31 December 2026
Format Options: You can either listen to the podcast or watch the video to complete the episode.
Requirement: Completing all 5 episodes in this series unlocks a final AI quiz.
Key Takeaways
The module frames Artificial Intelligence as a powerful tool to support public officers, drawing a parallel to how email and computers previously transformed workplace efficiency without replacing human roles.
Working alongside AI allows MOE staff to:
Boost Efficiency: Delegate routine tasks to AI to free up time for work requiring human judgment and empathy.
Enhance Service Delivery: Combine AI's data capabilities with an educator's deep understanding of the educational landscape.
Develop New Skills: Upskill technically while doubling down on uniquely human qualities like critical thinking and emotional intelligence.
Learning Objectives
By completing this bite, you will be able to:
Explain why AI matters to MOE staff and how it changes work dynamics via real-world examples.
Understand how AI works collaboratively with staff to enhance services rather than replacing the human element.
ThThe video from this mandatory learning bite highlights how Artificial Intelligence (AI) is being introduced as a collaborative tool for Ministry of Education (MOE) educators and staff.
Here is a summary of the core concepts presented in the module:
1. AI as a Workplace Co-Pilot
Rather than replacing human roles, AI is framed as an evolutionary step in workplace technology—much like how email and computers previously transformed administrative efficiency. It is designed to work hand-in-hand with staff to enhance public service delivery.
2. Key Benefits for MOE Staff
Automating Routine Tasks: By delegating repetitive, time-consuming chores to AI (such as processing data or generating summaries, as shown in the video's claymation visual), staff can free up time to focus on critical areas.
Amplifying Human Strengths: Freeing up administrative time allows educators and HQ staff to channel their energy into tasks that demand uniquely human qualities, such as empathy, critical thinking, emotional intelligence, and professional judgment.
Data-Driven Insights: Combining AI's data processing capabilities with an educator's deep understanding of Singapore's educational landscape allows for smarter, more optimized service delivery.
3. Skill Development
The module emphasizes that upskilling alongside AI involves a dual approach: technical familiarity with AI platforms and a renewed focus on reinforcing uniquely human interpersonal skills.
Note: This is Episode 1 of a 5-part series by the MOE Digital Excellence & Products Division (DXD). Completing the video or podcast for all 5 episodes is required to unlock the final AI quiz by the 31 December 2026 deadline.
2 Two
Here is a summary of the next mandatory learning bite, AI Ep 2: Breaking Down AI in Simple Terms, created by the MOE Digital Excellence & Products Division (DXD):
Overview
Topic: Breaking down the essentials of Artificial Intelligence and Generative AI (Gen AI).
Deadline: 31 December 2026
Requirement: Complete all 5 episodes in this series to unlock the final AI quiz.
Key Takeaways on Generative AI (Gen AI)
Definition: Gen AI creates entirely new content, including text, images, videos, music, and voices.
How it Works: Users describe what they want the AI to create by typing instructions into a chat dialogue box, known as a 'prompt'.
Capabilities: It functions as a writing and learning partner that can answer questions, quickly summarise complex information, and generate diverse ideas.
The Foundation: It is built on a massive amount of training data, allowing the model to identify patterns and similarities to produce its outputs.
Major Risks & Limitations
Hard to Distinguish: It can be difficult to tell the difference between human-made content and Gen AI output.
Misuse & Deepfakes: Gen AI can be misused to create "deepfakes" (realistic but fake images/videos) or generate text that sounds highly convincing but is factually incorrect.
Human Responsibility: Because of these risks, humans must always remain responsible for validating and fact-checking any content produced by AI.
Learning Objectives
By completing this episode, you will be able to:
Define AI and Generative AI, and identify an everyday example of its use.
Outline the three core steps of how AI learns and produces an output.
Identify the two biggest strengths of AI (speed and scale) and explain the major risk of relying blindly on training data.
The claymation video for Episode 2: Breaking Down AI in Simple Terms focuses on how AI actually processes information, using everyday analogies to make the tech accessible.
Here is a summary of the core concepts explained in the video:
1. The Three Core Steps of AI Learning
The video breaks down the machine learning process into a simple three-step pipeline:
Input (Data Collection): AI is fed massive amounts of training data (text, images, or code). The video uses the visual of a worker in a "coffee-fuelled data entry session" to illustrate the scale of data ingested.
Processing (Pattern Recognition): The AI model analyzes this data to find common threads, structures, and statistical probabilities.
Output (Generation): When you type a prompt into the chat dialogue box, the AI uses those learned patterns to predict and generate the most relevant response.
2. Generative AI (Gen AI) as a Creative Partner
The video highlights Gen AI's ability to create entirely new content from scratch—including text, images, videos, and music. It emphasizes using AI as a writing and learning partner to:
Answer complex questions instantly.
Summarize dense documents.
Brainstorm and generate diverse creative ideas.
3. Power vs. Pitfalls
The video reminds viewers of the balance between AI's strengths and its limitations:
The Strengths: Speed and scale. AI can process and output information faster than any human.
The Major Risk: AI relies entirely on its training data. If that data is biased or incorrect, the AI will confidently generate incorrect information.
The Deepfake Dilemma: Because AI-generated text, images, and videos look and sound incredibly realistic, they can easily be misused.
The Ultimate Takeaway: Because AI output can be highly convincing yet factually wrong, the video emphasizes that humans must always remain the final line of defense to validate, fact-check, and take responsibility for the final content.
3 Three
Based on the current page, here is a summary of AI Ep 3: Save These Gov-approved AI Tools in your Bookmarks, created by the MOE Digital Excellence & Products Division (DXD):
Overview
Topic: Using government-approved AI tools safely and effectively within the Singapore Public Service, with a focus on productivity and data security.
Deadline: 31 December 2026
Requirement: Complete all 5 episodes to unlock the final AI quiz.
Approved AI Tools by Data Classification
1. For Restricted / Sensitive Normal Data
These internal government tools are designed to handle more sensitive data securely:
Pair Chat Transcribe SmartCompose AIBots
2. For Official (Closed) / Non-Sensitive Data
Staff can log in using their MOE Google ICON email to access these commercial tools:
Gemini NotebookLM
Key Learning Framework: The 5-Part Prompting Framework
To get the maximum quality output from these AI tools, the episode introduces a structured framework for constructing clear and effective prompts:
Persona: Define the role the AI should adopt (e.g., "Act as an experienced primary school teacher").
Task: Clearly state what you need the AI to do.
Context: Provide background information or constraints.
Format: Specify how the output should look (e.g., bullet points, a table, or an email draft).
Tone: Dictate the style of communication (e.g., professional, encouraging, or concise).
The claymation video for Episode 3: Save These Gov-approved AI Tools in your Bookmarks highlights the practical implementation of AI tools within the Singapore Public Service, focusing heavily on matching the right tool to data classification and mastering the art of prompting.
Here is a summary of the core concepts shown in the video:
1. Matching Tools to Data Security
The video visualizes a clear boundary between different types of workplace data to ensure staff do not accidentally leak sensitive information:
For Restricted / Sensitive Normal Data: Use official government-secured platforms like
Pair Chat Transcribe SmartCompose AIBots For Official (Closed) / Non-Sensitive Data: Use commercial platforms like
Gemini NotebookLM
2. Masterclass in Prompting: The 5-Part Framework
To get the best possible results out of Gen AI tools, the video breaks down a structured framework for writing prompts, using a teacher planning a lesson as a primary example:
Persona: Tell the AI exactly who it is mimicking (e.g., "Act as a strict but encouraging primary school math teacher").
Task: Clearly define the objective (e.g., "Create a 5-question quiz on fractions").
Context: Provide background boundaries or student profiles (e.g., "For a class of 10-year-olds who are struggling with visual representations").
Format: Structure how you want the answer delivered (e.g., "Put it in a clear table with a separate answer key").
Tone: Set the vibe of the output (e.g., "Keep the language simple, engaging, and fun").
3. Expanding the AI Toolkit
The video rounds out by showcasing how AI isn't just a text box. It demonstrates how to leverage tools to:
Transcribe Audio: Turn meeting recordings or interviews into organized, actionable text summaries instantly.
Draft Correspondence: Use smart suggestions to reply to emails and messages with professional phrasing in seconds.
Build Custom Bots: Use AIBots to configure a chatbot tailored specifically to your unique workflow or department FAQs.
4 Four
Based on the page you are currently viewing, here is a summary of AI Ep 4: Here’s Why AI Can’t Replace Humans, created by the MOE Digital Excellence & Products Division (DXD):
Overview
Topic: Understanding the fundamental limitations of Artificial Intelligence and identifying where human expertise is non-negotiable.
Deadline: 31 December 2026
Requirement: Complete all 5 episodes in this series to unlock the final AI quiz. You can either listen to the podcast or watch the video to complete this bite.
The "Hard Boundaries" of AI
This episode flips the script on AI by focusing strictly on its flaws and boundaries, highlighting three major risks:
Hallucinations: AI can confidently generate incorrect, fabricated, or false information.
Frozen Data: AI models rely on fixed training data and lack real-time or up-to-the-minute knowledge unless explicitly updated.
Confidentiality Risks: Inputting sensitive information into unapproved AI tools poses a massive data security threat.
Because of these hard boundaries, human verification and critical judgment are absolute necessities.
A Practical Strategy: The "First Cut" Workflow
The module suggests a smart way to redesign your everyday professional workflow:
Let AI do the Heavy Lifting First: Use AI tools to generate "first-cut" drafts, brainstorm initial ideas, or clear baseline administrative tasks.
Save Your Energy for High-Stakes Work: By letting AI handle the first draft, you free up valuable time to focus on complex cases, ethical decisions, and meaningful, direct human interaction that truly requires empathy.
Learning Objectives
By completing this episode, you will be able to:
Identify the fundamental limitations of AI that make human verification essential.
Redesign your professional workflow by using AI for first-cut drafts while reserving your expertise for critical human-centric tasks.
The claymation video for Episode 4: Here’s Why AI Can’t Replace Humans visually reinforces the distinct boundaries between what machine learning can automate and what strictly requires human oversight.
The claymation video for Episode 4: Here’s Why AI Can’t Replace Humans visually reinforces the distinct boundaries between what machine learning can automate and what strictly requires human oversight.
Here is a summary of the core concepts presented in the video:
1. Confronting AI's "Hard Boundaries"
The video uses clear, everyday scenarios to illustrate three fatal flaws of AI that make blind reliance dangerous:
Hallucinations: The claymation characters show how AI can generate outputs that sound incredibly confident, professional, and grammatically perfect, yet are entirely fabricated and factually wrong.
Frozen Data: AI does not have lived experience or automatic real-time awareness. It is bound to its fixed historical training data, meaning it lacks contextual knowledge of sudden, current changes.
Confidentiality Risks: The video warns against plugging any sensitive, student-related, or internal school administration data into public, unapproved AI platforms, as it compromises data security.
2. Redesigning Work via the "First Cut" Strategy
Instead of fearing replacement or avoiding the tech entirely, the video champions a workflow model where humans and AI work in tandem:
The AI Phase (The First Cut): Let AI do the initial heavy lifting. Use it to generate rough outlines, draft basic templates, or brainstorm a wide list of starting ideas.
The Human Phase (The Final Polish): Save your energy for where it counts. Humans must act as the ultimate editors, taking that "first cut" and applying critical judgment, ethical evaluation, policy alignment, and fact-checking.
3. The Irreplaceable Human Element
The video concludes by reminding MOE staff that the most high-stakes, meaningful aspects of education can never be automated. Tasks involving complex student cases, ethical dilemmas, deep mentorship, and direct emotional empathy will always strictly require human decision-making.
5 Five
Based on the current page, here is a summary of the final mandatory learning bite, AI Ep 5: Your SAFE-ty Check Before Using AI Tools, created by the MOE Digital Excellence & Products Division (DXD):
Overview
Topic: Essential safety guidelines for using AI tools within the Singapore Public Service.
Framework: The S.A.F.E. four-part framework.
Deadline: 31 December 2026
Requirement: Completing this final episode unlocks the final AI quiz at the end of the series.
The S.A.F.E. Framework Breakdown
1. Safeguard Sensitive Information
Data Classification Matters: Always match your data type to the correct tool.
Internal Tools: Use government-secured platforms like PairChat for Restricted or Sensitive Normal data.
Commercial Tools: Only use official Open/Closed (non-sensitive) data when working with commercial AI tools like ChatGPT or Gemini.
2. Assess and Stay Accountable
Human Verification: You are legally and professionally responsible for AI outputs.
Fact-Check Everything: Mandatorily verify all high-impact decisions, cross-reference AI claims against official sources, and meticulously check any citations or references.
3. Follow Public Service Values
Purpose: Use AI explicitly to improve public service delivery and workplace efficiency.
Integrity: Never use AI to bypass personal accountability, cut ethical corners, or generate misleading/inappropriate content.
4. Educate Yourself and Engage Support
Stay Updated: Continuously check and align with your organization’s latest AI guidelines.
When in Doubt, Ask: If you are unsure about a tool's security or a dataset's sensitivity, immediately consult your Chief Information Officer (CIO) or ICT Manager.
Based on the current page, here is a summary of the final mandatory learning bite, AI Ep 5: Your SAFE-ty Check Before Using AI Tools, created by the MOE Digital Excellence & Products Division (DXD):
Overview
Topic: Essential safety guidelines for using AI tools within the Singapore Public Service.
Framework: The S.A.F.E. four-part framework.
Deadline: 31 December 2026
Requirement: Completing this final episode unlocks the final AI quiz at the end of the series.
The S.A.F.E. Framework Breakdown
1. Safeguard Sensitive Information
Data Classification Matters: Always match your data type to the correct tool.
Internal Tools: Use government-secured platforms like PairChat for Restricted or Sensitive Normal data.
Commercial Tools: Only use official Open/Closed (non-sensitive) data when working with commercial AI tools like ChatGPT or Gemini.
2. Assess and Stay Accountable
Human Verification: You are legally and professionally responsible for AI outputs.
Fact-Check Everything: Mandatorily verify all high-impact decisions, cross-reference AI claims against official sources, and meticulously check any citations or references.
3. Follow Public Service Values
Purpose: Use AI explicitly to improve public service delivery and workplace efficiency.
Integrity: Never use AI to bypass personal accountability, cut ethical corners, or generate misleading/inappropriate content.
4. Educate Yourself and Engage Support
Stay Updated: Continuously check and align with your organization’s latest AI guidelines.
When in Doubt, Ask: If you are unsure about a tool's security or a dataset's sensitivity, immediately consult your Chief Information Officer (CIO) or ICT Manager.
The claymation video for Episode 5: Your SAFE-ty Check Before Using AI Tools serves as the final review for the series, visually mapping out the S.A.F.E. protocol to ensure public officers handle AI tools securely and ethically.
Here is a summary of the core guidelines demonstrated in the video:
1. Safeguard Sensitive Information
The video uses clear visual boundaries to show how data classification must dictate tool choice:
Internal / Secured Data: If you are handling Restricted or Sensitive Normal public service data, you must only use official government-secured environments like PairChat.
Public / Non-Sensitive Data: Commercial tools like ChatGPT or Gemini are strictly reserved for Official (Open/Closed) non-sensitive data.
2. Assess and Stay Accountable
The video emphasizes that AI is not a legal shield.
The "Human-in-the-Loop" Rule: You are personally and professionally accountable for any material or decisions generated by AI.
Mandatory Fact-Checking: You must verify any high-impact claims against trusted official sources and carefully check that any citations or references provided by the AI actually exist and are accurate.
3. Follow Public Service Values
The animation highlights the ethical use of technology:
AI should be used transparently to optimize workflow efficiency and improve public service delivery.
It must never be used to bypass personal responsibility, cut ethical corners, or generate misleading, biased, or inappropriate content.
4. Educate Yourself and Engage Support
Because AI technology and governance policies change rapidly, the video concludes with an emphasis on continuous learning:
Stay updated on the latest MOE and Public Service AI guidelines.
Reach Out: If you are ever uncertain about the data classification of a file or the security of a new AI tool, immediately consult your Chief Information Officer (CIO) or ICT Manager before proceeding.
The claymation video for Episode 5: Your SAFE-ty Check Before Using AI Tools serves as the final review for the series, visually mapping out the S.A.F.E. protocol to ensure public officers handle AI tools securely and ethically.
Here is a summary of the core guidelines demonstrated in the video:
1. Safeguard Sensitive Information
The video uses clear visual boundaries to show how data classification must dictate tool choice:
Internal / Secured Data: If you are handling Restricted or Sensitive Normal public service data, you must only use official government-secured environments like PairChat.
Public / Non-Sensitive Data: Commercial tools like ChatGPT or Gemini are strictly reserved for Official (Open/Closed) non-sensitive data.
2. Assess and Stay Accountable
The video emphasizes that AI is not a legal shield.
The "Human-in-the-Loop" Rule: You are personally and professionally accountable for any material or decisions generated by AI.
Mandatory Fact-Checking: You must verify any high-impact claims against trusted official sources and carefully check that any citations or references provided by the AI actually exist and are accurate.
3. Follow Public Service Values
The animation highlights the ethical use of technology:
AI should be used transparently to optimize workflow efficiency and improve public service delivery.
It must never be used to bypass personal responsibility, cut ethical corners, or generate misleading, biased, or inappropriate content.
4. Educate Yourself and Engage Support
Because AI technology and governance policies change rapidly, the video concludes with an emphasis on continuous learning:
Stay updated on the latest MOE and Public Service AI guidelines.
Reach Out: If you are ever uncertain about the data classification of a file or the security of a new AI tool, immediately consult your Chief Information Officer (CIO) or ICT Manager before proceeding.
6 Six
Based on the page you are currently viewing on glow, here is a summary of the final requirement for the AI Literacy series:
Overview: AI Literacy Quiz
This is the final mandatory assessment created by the MOE Digital Excellence & Products Division (DXD) to conclude the AI Literacy training.
Status: Required / Mandatory
Deadline: 31 December 2026
Completion Criteria
Prerequisite: You must fully complete Episodes 1 to 5 of the AI Literacy series to unlock this quiz.
Passing Score: A minimum score of 80% is required on the quiz to officially mark the completion of your mandatory AI Literacy series.
1
Here is why, based on the Gov-approved tools covered in Episode 3:
AIBots = custom chatbots (allows you to customize AI bots for specific use cases)
SmartCompose = drafts documents (helps draft faster, professional responses)
Transcribe = speech to text (transforms audio/speech conversations into text)
Pair = everyday LLM assistant (acts as the secured ChatGPT for the Singapore Government)
Why these two are the most effective:
They follow the 5-Part Prompting Framework covered in Episode 3 by providing clear instructions, context, roles, and constraints:
Prompt B is highly effective because it defines a Persona ("As a data analyst") and gives clear context and specific parameters for what to look out for ("recurring suggestions and concerns").
Prompt D is highly effective because it clearly specifies the exact Format it wants the output in ("Top 3 concerns, % of positive vs negative responses, Key suggestions"), making it highly actionable for the AI.
Why the others are less effective:
Prompt A is too vague ("tell me what people are thinking"), which leads to broad, generic AI responses.
Prompt C changes the actual sentiment of the feedback ("make it... positive"), which defeats the purpose of trying to identify genuine insights and concerns from the citizens.
3
Why this is correct:
Throughout the series, these tasks were highlighted as excellent examples of how AI can handle the "first-cut" and routine administrative workload to free up your time:
A refers to the capabilities of tools like Transcribe and Pair Chat to handle meeting minutes.
B and C highlight the core strengths of Gen AI—speed and scale—especially when using tools like NotebookLM or Gemini to summarize and find information across multiple documents.
D is the exact use case for SmartCompose and text assistants to help you draft routine correspondence quickly.
Why this is correct:
All five statements represent real risks and limitations of AI that require the "human-in-the-loop" to check, verify, and take ultimate accountability for:
Statement A (Hallucinations): As highlighted in Episode 4, AI can generate beautifully phrased but entirely fake information.
Statement B (Frozen Data): AI is restricted to its training cut-off date and lacks real-time awareness of sudden changes.
Statement C (Data Classification): AI doesn't inherently understand the real-world stakes of sensitive government data, which is why humans must safeguard what gets typed into it (Episode 5).
Statement D (Legal/Copyright): Content generated by commercial AI tools can run into compliance and intellectual property issues.
Statement E (Lack of Context): AI excels at patterns but lacks genuine human nuance, empathy, and organizational context.
Why this is correct:
These three practices directly align with the 5-Part Prompting Framework (Persona, Task, Context, Format, Tone) covered in Episode 3:
Practice B (Break down complex tasks): AI performs significantly better when you guide it through step-by-step instructions rather than dumping a massive, complicated request all at once.
Practice C (Be specific with tasks): Vague instructions lead to generic, unhelpful answers. Clear boundaries give you actionable results.
Practice E (Specify format and tone): Telling the AI exactly how to present the information (e.g., in a table, bullet points) and what vibe to use (e.g., professional, encouraging) saves you editing time.
Why the others are incorrect:
Practice A is wrong because giving the AI too little information removes the necessary context it needs to generate a relevant answer.
Practice D is wrong because "vague prompts" usually result in AI hallucinations or off-topic outputs rather than helpful "creative freedom."
Why this is correct:
Statement A (Appropriate): Episode 4 highlighted the "First Cut" strategy, where using AI to draft routine queries frees you up for more complex work.
Statement B (Appropriate): Summarizing large volumes of information plays directly to AI's strengths of speed and scale (Episode 2).
Statement C (Appropriate): Human expertise is non-negotiable for tasks that require ethical judgment, nuanced decision-making, and emotional intelligence.
Why the others are incorrect:
Statement D (Inappropriate): Public service values and safety guidelines state that humans must remain accountable for final decisions; we cannot delegate sensitive policy decisions to AI.
Statement E (Inappropriate): AI lacks genuine empathy and human context, making it completely unsuitable for leading complex social discussions.
Why this is correct:
The bullet points listed directly in the question text already provide the completely accurate definitions for each term based on your training module:
Lack of Context is perfectly matched with the inability to retain/access previous conversation info.
Confidentiality is perfectly matched with the risk of data being stored and used for future training.
Cut-off Date is perfectly matched with only providing information up to its last training date.
Copyright is perfectly matched with the risk of incorporating protected content without permission.
Bias is perfectly matched with reflecting stereotypes from its training data.
Since the list in the prompt is already completely accurate, option C is the correct choice.
8
Why this is correct:
To mitigate the risks of AI (such as hallucinations, bias, and incorrect information), a "human-in-the-loop" must actively verify the output before it is used or passed along:
Step A, B, and D directly address AI's tendency to confidently fabricate facts, calculations, and citations.
Step C ensures that any hidden stereotypes or skewed perspectives from the AI's training data are caught.
Why Step E is incorrect:
Step E goes completely against the core principle of Accountability covered in Episode 5. Public officers are personally and professionally responsible for the work they submit. Presenting an AI draft to your supervisor without any human review or verification is inappropriate and risky.
Why this is correct:
These four statements make up the S.A.F.E. framework for responsible AI use introduced in Episode 5:
Safeguard sensitive information (Statement A)
Assess and stay accountable (Statement C)
Follow public service values (Statement D)
Educate yourself and engage support (Statement E)
Why the other statements are incorrect:
Statement B ("Endure AI's limitations...") is not a guideline; the framework teaches you to mitigate and check for limitations, not just tolerate mistakes.
Statement F ("Store all information on AI...") violates data security protocols, as sensitive government data should never be casually stored on unapproved AI systems.
Why this is correct:
AI is the broad umbrella term for systems that mimic human intelligence to perform a wide variety of tasks—such as recognizing speech, analyzing patterns, or processing images.
Generative AI is a specific branch (or subset) of AI that focuses on creating entirely new content (like text, images, or audio) based on the patterns it has learned from its training data.
Why the other options are incorrect:
A mixes up the definitions; AI as a whole handles things like predictions, not just Gen AI.
B is incorrect because Gen AI is distinct in its ability to create brand-new creative outputs.
D is incorrect because many other types of non-generative AI (like predictive models or classification systems) also learn from data.
Why this is the correct lifecycle:
AI is trained on large amounts of data: Before an AI can do anything, it must first be fed a massive dataset (like books, articles, or code) so it has material to learn from.
AI learns patterns from the data: During the training phase, the system analyzes the dataset to detect relationships, structures, and recurring patterns between words or numbers.
When you ask it something, it uses the patterns to give relevant answers: Once trained, the AI applies those learned patterns to predict and generate the most logical response to your prompt.
Why this is correct:
As emphasized in the S.A.F.E. guidelines (Episode 5) and the tool-matching framework (Episode 3):
Public commercial AI tools (like the free/public versions of ChatGPT or Gemini) do not have government-grade security protocols.
Therefore, you are only permitted to upload or type Official (Open/Closed) and completely Non-Sensitive information into them.
Why the other options are incorrect:
B, C, and D all involve data that is restricted, sensitive, or confidential. Putting any of this data into a commercial AI tool constitutes a serious security breach. For Restricted or Sensitive Normal data, you must use government-secured tools like Pair Chat.
No comments:
Post a Comment