I’ll cover how to choose a set of tools that work together and work for different people or groups. It isn’t as straightforward as you might expect, particularly for groups such as After School Clubs. I’ll only be looking at the tools used at the Innovation Centre, although alternatives may be available. Note: I’ve listed how we use the tools not how they can be use. For example X AI can the used for Graphics but we only use the Text Generation functions.
List of tools used at the Innovation Centre
Digital Camera (Much of the art starts from a photo)
Adobe Elements (Digital Art and Video)
JWildFire (Fractals)
FlameLet (iPad Fractals)
ProCreate (iPad Digital Art)
Paper (iPad simple sketching/painting)
Scratch (coding, animation and robotic control for beginners and children)
Scratch – Raspberry Pi (designed for coding, animation and physical computing )
Machine Learning for Kids (Scratch, plus AI machine learning)
Google’s Teachable Machines (Simple AI Learning)
Raise Playground (Scratch AI machine learning by MIT)
MakeBlock (Chinese alternative to Scratch with robotics and computer control)
MakeCode (Scratch style blocks for robotics – Microbit, Lego Ev3)
Mind+ (Chinese alternative to Code with a focus on electronic components)
Scratch JR (simple scratch alternative for younger children)
WeDo (simple Lego robotics for younger children)
WeDo app (Educational Environment for WeDo. Runs on a tablet)
Lego Boost (next step up from WeDo)
Lego Boost app (Educational Environment for Boost. Runs on a tablet)
Ev3 (Lego Robotics)
Ev3 apps (Lego MindStorm ,original coding; Classroom, block based and Python)
Raspberry Pi (Large range of single board computers with pins to control electronics and robotics, plus AI)
Arduino (Wide range of Micro Processors for controlling robotics and electronics, and Physical AI)
Microbit (A wide range of learning kits – an alternative to the Arduino for children)
3D Printer (used to create models and enclosures for electronics and robotics)
Monitoring and control components (add on to Pi, Arduino and Microbit. Including physical AI)
Robotics (Range of kit for use with Arduinos, Raspberry Pis and Microbits)
Electronics (Range of kits and items for use with Arduino, Raspberry P1 and Microbits)
3D printer
Processing/p5.js (computer language ideal for art and robotics)
Python (the most common computer language, ideal for general purpose and AI)
Micro Python (language designed for micro processors and robotics)
Microsoft AI (AI Text and image)
Google AI (AI Text)
ChatGPT (AI Text)
X AI (Text)
Perplexity AI (AI Text for science and technology)
DeepDream Generator (AI Image)
Suno (AI Music)
Adobe Express (new – Digital Art and AI for beginners and children)
Assumptions
The aim is to try out new ideas, integrate different elements and do something innovative. We are also looking to gain skills in Digital Art, Robotics and AI plus other STEAM disciplines.
For Younger Children and beginners
Children 7+
Use Lego WeDo to build robots and program them. The Lego WeDo app provides lots of educational content and superb build and coding instructions.
Use Scratch JR to create simple animations.
Once you are familiar with the basics, create your own backgrounds and characters.
The final step is to coordinate the animation with the robot (this is dome manually)
Older Children could start to redevelop a project using Scratch, rather than using the WeDo App and Scratch JR. This will allow the Scratch animation to control the robot.
Children 9+
We can now use Scratch alongside Lego WeDo, Boost or EV3.
We can use Generative AI to help in creating interesting backgrounds and characters.
We can use a MicroBit rather than Lego but with a very limited set of functions.
For older children or those more experienced, we can add AI by running Scratch from machinelearningforkids.
Children 11+
Lots more opportunities are available.
Explore fractal graphics with JWildFire and use advanced AI Graphics with a range of AI Art Generators. Use the fractals as styles for the AI Art.
Use AI to create or supporting stories
Use a Raspberry Pi to control electronic components and also run Scratch. Try creating a project that runs on the Raspberry Pi. Create an integrated project that uses Scratch Animation and a Robot Car controlled by the Raspberry Pi. It is possible to integrate them but it isn’t easy and required a sound knowledge of GPIO.
Look at using an Arduino for robotics.
More Experienced Learners
There is a huge range of possibilities. Try a project that integrates Art, Robotics and AI.
You can use Scratch for Art and Animation, or replace it with Processing/ p5.js which is designed for artwork and is ideal for robotics when combined with an Arduino.
Try out all the AI capabilities of Deepdream Generator. Start to combine this with other generators and Fractal Graphics Generators. Start to use AI as a smaller component when creating artwork.
Build sophisticated robots. Lego Mindstorm is great for building complex prototypes. However you can mix and match components e.g. Lego, Meccano and robotics kits. MicroBits, Arduino or Raspberry Pi can be used for robotics, or combined to create individual robots or swarms.
Look at using Physical AI and Edge AI with either a Raspberry Pi, Arduino or Microbit. Create Smart AI Robotics. Integrate these with AI Art.
Try out different programming languages such as Python, Micro Python (designed for micro processors), Processing, Arduino C, Java Script.
Create integrated projects that cover Art, Robotics and AI, and also include stories and music. Build these projects around a theme. Use Generative AI to investigate the theme e.g. Climate Change, The future role of AI.
Build AI Projects that cover a wide range of AI, including Generative AI, AI Algorithms such as Random Forests, Physical AI and Edge AI.
Problem Areas
Integration
The issue when creating integrated projects, that encourage innovation, is that manufactures often make tools and products that are difficult to integrate. An example of the issue is using an AI vision sensor on a MicroBit robot that integrates with a Scratch animation, which uses AI gesture control. All designed with children and beginners in mind but difficult to integrate. For anyone reading this who knows how to do it, its relatively simple but you need design skills beyond the basic level.
Although Scratch is widely used by children and adults to make simple animations, and Lego is the top building tool, integrating robotics with Lego is getting harder. While Lego Spike uses a Scratch-like language, it doesn’t work with Scratch. The robotics compatibility with Scratch varies by environment, such as Standard Scratch, MLforkids, Raise Playground, Mind+, and MakeBlock. Lego Ev3, Boost, and WeDo are compatible with most environments. Microbits, Arduinos, and Raspberry Pi have limited support or functionality. Chinese Scratch environments like MakeBlock and Mind+ offer better integration but can be more challenging to use.
Creating integrated projects with AI for Physical or Edge Computing needs good educational resources. Raspberry Pi Foundation, Lego, Arduino, Scratch, Machine Learning For Kids, MIT, and edX offer great materials. However, training from Chinese Scratch providers like MakeBlock (mBlock 5) and Mind+ is not very useful for English users.
At the Innovation Centre we recommend tools that integrate and work well together.
Alternative Approach to Integration
It is often possible to find ways to connect different systems. However, another way is to integrate when you can and combine when it’s difficult to connect.
For example, using the new Lego Spike with Scratch (or Scratch AI) shows this approach. Here, the animation can show when to start the robot or begin an action. This practical method boosts innovation skills and creativity. We can mix apps or robots that work on various devices like laptops, phones, and tablets, allowing larger groups to work together.
Generative AI has an issue with Bias
Having used Generative AI for a number of years and monitored the Bias in both Image Generation, stories and answers to research questions, things have changed considerably over time.
It is obvious if you monitor a wide range of Generative AI tools that they all have Bias and the type of bias is changing constantly.
In the early days of AI images and story styles were predominantly based on the data used to train the AI tool, which led to some bias.
Later it became obvious that bias was being designed out or in. From the outside this looked like trial and error.
A number of specific generators also tend to adjust bias due to “likes”
Generators know what you have used in the past and base responses on your history. Your history is therefore adding Bias, which may not be helpful.
Some Generators allow you to use conversations where each request is seen as part of the conversation.
Understanding bias is essential for checking the validity of answers and for framing questions or conversations to achieve desired results. This applies to both discussions and Generative Art. Bias can be difficult to recognize and overcome, requiring practice to master.
Generative Art
Using Generative Art to create or enhance images requires time and practice. Unfortunately, there isn’t much information on how these tools work. For example, increasing detail in your description may help or hinder the results, depending on the Generative Model you choose. Some models rely on keywords, while others can understand sentences. Deepdream Generator offers various models and features, producing different results even with the same prompt.
Southport Innovation Centre 