First Steps in AI and Data Analysis

Artificial intelligence (AI) and machine learning (ML) are the technologies that are likely to have the greatest day-to-day impact on young people’s lives across the world over the next five to ten years. Starting to teach them about it and how to use the technology responsibly, from an early age is very important.

The development of AI and ML technology has happened so quickly that school curriculums haven’t had time to adapt and there is very little specific guidance, or resources, published to support teachers.

Allowing students opportunities to explore AI and ML applications in a safe space with clear guidance will let them develop an understanding of the different contexts in which the technology can be used and the impact it can have. There are already many misconceptions about AI and how it might develop in the future. Misinformation and ‘fake news‘ are prevelant across the internet and it is crucial to ensure students are properly educated about this and are prepared for the truth.

AI and ML technology are moving so fast (4 ChatGPT versions in six months to January 2024), and new tools are constantly being developed and released. There are already freely available online AI applications for creating text, images, music, video, translating languages and so on, and the opportunities are only going to expand over the next few years.

Many chatbots such as ChatGPT and Google Gemini require you to set up a free account. You can then input a prompt such as, “Write a short definition of Artificial Intelligence to explain it to a 10-year-old child.” Examples of outputs are given in the PPT presentation.

ChatGPT can currently be used in several languages, including English, Spanish, French, German, Portuguese, Italian, Dutch, Russian, Arabic, and Chinese.*

Google Gemini supports over 40 languages.*

*Information correct as of February 2024.

How close were the students’ ideas to the definitions written by the AI chatbots?

Ask students to identify current uses of AI & ML that they think already affect their daily lives. Make a list on the board and compare it to the list in the PPT file.

Show students this short video to explain the difference between AI and ML:

Ask students to discuss possible future applications of AI and ML and how they think those could impact their lives – both positively and negatively.

Written by Google Gemini

Ever wondered why the cereal you loved suddenly vanished from the shelves, or why you get coupons for carrot sticks just when you were thinking about them? It's not magic, but it's pretty close! Supermarkets have sneaky little helpers called robots, powered by Artificial Intelligence (AI), that use information from your loyalty card to predict what you'll buy and make the store better for everyone.
Imagine your loyalty card as a chatty chipmunk, whispering what you buy to a giant, friendly robot brain. This brain collects whispers from all the chipmunks, then uses its AI muscles to understand what everyone loves (like mountains of ice cream!) and what's not so popular (sorry, brussels sprouts).

Here are some cool things these robot brains can do with your whispers:
Treasure Hunt: Ever noticed how milk and bread are always at the back of the store? That's not a coincidence! The robot brain figured out most people buy these things every time, so putting them far away encourages you to explore and buy other stuff. It's like a supermarket treasure hunt!
Coupon Caper: Remember those carrot coupons? The robot brain noticed you bought broccoli last week, so it guessed you might like other healthy snacks. It's like having a personal shopper whispering deals in your ear, helping you save money and eat your veggies!
Shelf Shuffle: The robot brain also sees which products disappear quickly from the shelves and which ones gather dust. This helps the supermarket stock the popular stuff more often and maybe even introduce new snacks you'll love based on what other chipmunks are buying.
Happy Shopper Party: But all this whispering isn't just about selling more stuff. The robot brain wants you to have a happy shopping trip! It remembers when the checkout lines get long and tells the supermarket to open more registers. It even sees which products are confusing or hard to find and tells them to put them in better places.

Of course, you might be wondering, "Isn't this a bit creepy?" It's true, super shiny robots collecting our shopping data sounds a bit like something from a movie. But remember, the robot brain doesn't know who you are, just what you buy. It's like your whispers are anonymous, helping make the supermarket a friendlier, tastier place for everyone.
So next time you swipe your loyalty card, remember the friendly robot brain behind the scenes, working hard to make your shopping trip super! Think of it as a high-tech game of "I Spy" with benefits – everyone wins!
 

• Does your family use loyalty cards?
• Did you know how supermarkets use the data?
• Do you think this is a positive or negative way to use data about your shopping habits?

The following short video will explain to older students how data analysis with AI can be done. The benefit of AI is that scaling it up to analyse data for millions of customers can be done very easily and quickly.

We need computers to be able to analyse huge data sets and identify patterns – which would take humans an impossibly long time to do. This ability to handle huge amounts of data is something computers are excellent at and it is the starting point for AI and ML.

Two examples of handling huge amounts of data are the MET Office (weather service) in the UK and the European Centre for Medium-Range Weather Forecasts (ECMWF). Both organisations use some of the most powerful computers in the world to process hundreds of billions of data points every day to forecast the weather across Europe, and beyond. The supercomputers used by both organisations are capable of performing 1000s of trillions of calculations every second!

Both the MET Office and ECMWF are now starting to use AI and ML to process and analyse the data they collect, identify patterns and help forecast the weather. The main advantage of this is the amount of data that can be processed in a short space of time, efficiently and cheaply. This article – ‘Capturing uncertainty in forecasts based on AI’ explores the advantages of using AI systems in weather forecasting.

These two videos examine the improvements AI can make to weather forecasting and how that can improve safety and ultimately save lives.

Can AI help weather forecasting save lives?

Machine Learning for Weather Forecasts at ECMWF

Write a 300-word* story, suitable for 11-year-olds⁺, about an adventurous aardvark who sets off to find some hidden treasure.

*Always specify the word length – short texts are written quicker.

⁺Always specify age suitability. This will ensure that nothing inappropriate is included and that the vocabulary level is appropriate.

Write a 300-word article, suitable for 10-year-olds, about endangered species in Europe.

Write a 250-word biography of scientist Marie Curie suitable for 12-year-olds.

See the ‘Large Language Model Examples’ document (word file and PDF) for more generative AI examples.

• Ask multiple students to input exactly the same prompt and see what output they get.
• Ask students to edit their prompt to be more specific and see how that changes the output.
• Try prompts that include specific types of writing; for example, a diary entry, a story in the first person or a factual piece including bullet points and a glossary.

Demonstrate to students how Google Translate can translate over 130 languages. It can translate from typed words, spoken words and from images.

• Give students a text in a different language, Mandarin for example, and ask them to translate it.
• Pair students up and challenge them to have a conversation in a different language using two devices.

Students try using Canva or Dall.E to create their own images by inputting text descriptions. Let students explore creating different image types and what happens as they edit their text descriptions. See ‘AI Generated Image Examples’ (PPT and PDF) for examples of prompts, the images AI can create, and some of the limitations of AI image generation.

Discuss the ethical issues around AI image generation – scraping, copyright, ownership and accountability. This article is an excellent starting point. This video explains how AI image generation works:

More links and resources regarding ethical, social and ecological implications of artificial intelligence can be found in the background article to the AI in STEM Education project: AI in Education: Theory and Practice.

Students can also try Teachable Machine – an online tool where they can train a computer to recognise shapes, colours, movements and sounds and then sort through data to identify them.

In our teaching resource “EcoKids Teach AI” targeted to younger students, a Teachable Machine model is trained to classify recyclable waste.

The Primary Science Teaching Trust’s ‘A Scientist just like me’ programme has compiled short profiles of scientists working with AI and machine learning.

Here you can download the PDFs directly:
Artificial Intelligence Researcher – Krishna Mooroogen
Data Scientist – Dr. Fiona Reid
Machine Learning Scientist – Kaushik Hatti
Software Engineer – Rebecca Harding

Activity 1–3:

PPT: First Steps in AI and Data Analysis

Activity 2:

PPT: Supermarket Receipts

Word: Supermarket secrets

PDF: Supermarket secrets

Activity 3:

Dice Data spreadsheet

Activity 4:

Word: Large Language Model Examples

PDF: Large Language Model Examples

Activity 5:

PPT: AI Generated image examples

PDF: AI Generated image examples