Cheat Sheet!
A) ChatGPT Prompt Pack
https://academy.openai.com/public/tags/prompt-packs
Ready-to-use prompt examples to help you get the most out of ChatGPT
B) Google AI Boost Bites
https://www.cloudskillsboost.google/paths/2480
A series of concise, impactful video lessons designed to elevate your expertise in the dynamic world of artificial intelligence.
C) Langchain Hub (for prompts)
https://smith.langchain.com/hub
Centralized platform designed for the management, sharing, and discovery of various artifacts used in building applications with LangChain, particularly those related to large language models (LLMs).
D) Perplexity at Work
A Guide to Getting More Done
https://r2cdn.perplexity.ai/pdf/pplx-at-work.pdf
E) CO-STAR Framework for Effective Prompts
- Context: The background information or scenario relevant to the task, helping the AI understand the situation.
- Objective: A clear definition of the specific goal or task you want the AI to perform.
- Style: The desired writing style for the response, such as technical, creative, or formal.
- Tone: The emotional or attitudinal quality of the response, like enthusiastic, cautious, or professional.
- Audience: Identifying the intended readers or users of the response helps tailor the language and content.
- Response: Specifying the desired format of the output, such as a list, table, or paragraph.
Three golden rules to keep in mind as you explore the world of AI.
1. Check the Facts, Hor! (Your AI Can Sometimes "Gostan")
Ever heard of AI "hallucinations"? It's a fancy term for when an AI confidently makes stuff up. It can write something that looks totally believable but is actually wrong. It’s not trying to lie, it’s just trying to predict the next word and sometimes gets it wrong.
So, if you're asking for important information—like medical symptoms, financial advice, or legal stuff—don't just take the AI's word for it. Always double-check with trusted sources you know are legit, like government websites (MOH, CPF), academic journals, or mainstream news outlets.
2. Don't Share Your NRIC! (And Other Secrets)
This one is simple but super important. Treat public AI chatbots like a conversation in a crowded coffee shop. You wouldn't shout your personal details there, right?
Never, ever type sensitive personal data into a public AI. This includes:
Your NRIC or FIN number
Bank account or credit card details
Your home address
Confidential information from your workplace
Keeping your private data private is the easiest way to protect yourself online.
3. Learn to Spot the Fake
With AI, creating fake videos and audio clips (aka "deepfakes") is getting easier. We've all seen those weird videos that look almost real. They're a growing concern, especially for spreading misinformation.
Here’s a quick checklist to help you become a deepfake detective:
Look at the eyes: Does the person blink weirdly or not at all?
Watch the face: Are facial movements unnatural or do emotions seem "off"?
Check the background: Do you see strange lighting, blurry shapes, or distorted lines?
Listen closely: Does the audio sound robotic or out of sync with their mouth?
When you see a sensational video, especially on WhatsApp or TikTok, take a moment and be skeptical. A little bit of caution goes a long way!
No comments:
Post a Comment