The AI Family Tree: 4 Types of AI Your Kids Use Every Day

AI isn't one thing — it's a family of different technologies, each specialised for a different job. In the AI Adventures course, Parikshet introduces these as four characters: Seeko, Chatty, Arty, and Gamer. Here's the full breakdown.

Type 1: Recommendation AI ("Seeko the Suggester")

What it does: Predicts what you'll want to see, buy, listen to, or watch next — based on your history and what similar users did.

Examples kids use:

• YouTube autoplay and sidebar suggestions
• Netflix "Because you watched..." rows
• Spotify Discover Weekly
• Amazon "Customers also bought..."
• TikTok For You page

How it works: It compares you to millions of other users. If people similar to you (based on watch/listen/buy history) enjoyed something, it predicts you will too. It doesn't understand the content — it just understands patterns of human behaviour.

Why it matters: Recommendation AI shapes what information children (and adults) are exposed to. Understanding that a recommendation is an algorithm's prediction — not a neutral suggestion — is crucial media literacy.

Type 2: Language AI ("Chatty the Talker")

What it does: Understands and generates human language — answering questions, writing text, translating, summarising.

Examples kids use:

• ChatGPT, Claude, Google Gemini
• Google Translate
• Autocorrect and predictive text
• Voice assistants (Siri, Alexa, Google Assistant)
• Email auto-reply suggestions

How it works: Language AI (specifically Large Language Models) was trained on billions of pages of text. It learned statistical patterns in language — which words and ideas tend to appear together, how sentences are structured, how arguments are made — and uses those patterns to generate text that fits the context.

Key limitation: Language AI doesn't "know" things the way humans know things. It generates statistically plausible text. That's why it can sound confident while being wrong.

Type 3: Vision AI ("Arty the Artist")

What it does: Understands and generates images — recognising objects, faces, scenes, and creating new visual content.

Examples kids use:

• Face unlock on phones and tablets
• Google Photos "search by what's in the photo"
• Instagram and Snapchat face filters
• AI image generators (DALL-E, Canva AI, Adobe Firefly)
• Google Lens (point your camera at something to identify it)

How it works: Trained on millions of labelled images, the AI learned to recognise visual features — edges, shapes, textures, and how they combine into objects and faces. Image generation AI works in reverse: given a text description, it generates pixels that match the description.

Type 4: Game AI ("Gamer the Thinker")

What it does: Controls characters in games, makes decisions in response to player actions, creates dynamic game behaviour.

Examples kids use:

• Enemy characters in video games that respond to your strategy
• Chess or chess-style game opponents
• Sports game opponents that adapt to your playing style
• NPCs (non-player characters) in open-world games that have conversations

How it works: Game AI ranges from simple decision trees ("if player is near, attack") to sophisticated learning systems that observe how players play and adapt. Modern AAA games increasingly use machine learning to make AI opponents more realistic and challenging.

The activity: Match the AI

This is Lesson 1.3 in AI Adventures. Try it with your child: list 10 apps or devices in your home and classify each as Seeko (recommendation), Chatty (language), Arty (vision), or Gamer (game). Some will use more than one type — a modern phone camera uses Vision AI to recognise faces AND Language AI to generate captions. That's the fun of the exercise.