IntegrAuth Academy
Identity, security and practical AI — free, from zero. Byte-sized lessons, real diagrams, hands-on labs, zero fluff. Two programs, each with its own exam and certificate.
English is hard? Your browser can translate this page into Tamil, Telugu, Kannada, Malayalam or Hindi. Tech words stay in English so you learn the real names.
How?
- Chrome on phone: ⋮ menu → Translate
- Chrome on computer: right-click → Translate to …
- Edge: the translate icon in the address bar
Your first AI chat
Priya has never tried AI. Her manager says, "Just ask it something." Your goal: have a first real chat, and learn what never to type in.
🧪 Interactive lab — enable JavaScript to play with this one.
- An app you talk to in plain words is a Chat assistant. An AI model inside it writes the answers.
- Say what you want in full sentences. Then steer with follow-ups, like "cheaper."
- Start a new chat for each new topic.
- Two answers disagree on a fact? Check a real source.
- Never paste passwords, codes, card or ID numbers, or other people's private details.
Each chat starts fresh
The AI reads only the messages in this chat. All the text it can read at once is its Context window. A new chat starts blank, like a fresh page in a notebook. Only saved instructions and memories come along. An old topic can confuse a new answer. Very long chats also get slower and worse.
Temporary chats and shared links
A Temporary chat is not saved in your history or memory. Most apps also do not use it to train the AI. Use it for a question you don't want to keep. But it is not invisible. The company may still keep a copy for a short time.
The Share button makes a Shared link. Anyone with the link can read the whole chat.
Everything you type goes to the company's computers. So what is safe to paste? Sort 12 examples.
🧪 Interactive lab — enable JavaScript to play with this one.
Go deeper: the model and the app
The model is a program. It learned from a huge amount of text. Now it is very good at one thing. It guesses which words are likely to come next. It writes each answer a small piece at a time. To learn more, read what an AI model is.
The app around the model adds everything else. That means the chat window, your past chats, files and photos, web search, voice, memory and safety rules. Think of a car. The model is the engine. The app is the rest of the car. You are the driver. The car goes where you steer it, even the wrong way.
Go deeper: which assistant to use
As of September 2026, well-known assistants include ChatGPT, Claude, Gemini, Microsoft Copilot, Meta AI, Le Chat and DeepSeek. They are in no order. This is not advice to pick one. All of them work on a website. Most also have phone apps. Some live inside apps you already use. Gemini is in Gmail, Copilot is in Windows, and Meta AI is in WhatsApp.
Each one has a free version that is good enough to learn on. Paid plans give you more use, the strongest models and extras. A bigger difference is where your words go. Read the company's Privacy policy. It says how long chats are kept, and if the company trains AI on them. It also says where your data is stored. At work, use the assistant your employer approved. See safe AI at work.
Go deeper: the buttons worth knowing
Names and places change between apps. But as of September 2026, the big assistants all have some version of these.
- ✏️ Edit an earlier message. The AI answers again from that point. The old version drops out of the chat.
- 🔁 Regenerate gives a new answer to the same message. The model picks each word partly by chance, so answers differ. Learn why answers change.
- 📎 Attach a PDF, a screenshot or a photo. Then ask about it, like "Translate this menu."
- 🎙️ Voice mode lets you talk out loud and hear the reply.
- 📁 Projects and Custom instructions save things you would repeat, like "I'm a nurse, keep answers short."
- 🧠 Memory saves notes about you, like your job, for new chats. You can usually see, edit or delete them. You can also turn memory off.
Go deeper: sharing and privacy
In July 2025, thousands of shared ChatGPT chats showed up in Google search. Their owners had ticked a box to make them "discoverable." Some chats had names and personal stories in them. OpenAI removed that option within days. So treat every shared link as a public page. Read a chat again before you share it. You can delete shared links later in the app's settings.
Your own health, money or legal questions are your choice. Decide on purpose, and think about a temporary chat. You can often swap private details for a Placeholder, like "[my landlord]" or "account ending [XXXX]." The answer is usually just as good. Priya does this with a letter from her insurance company. She covers her policy number with her thumb before she takes the photo.
An AI sounds just as sure when it is wrong. The next lesson shows where AI goes wrong.