​ 'LEGO® Bots': AI Builders
Techno-workshop by NeoBreizh Studio
​
Imagine a world where every child and adult can become an artificial intelligence architect through LEGO®! With 'LEGO® Bots: The AI Builders', we transform the complexity of AI into a creative and accessible adventure.
​
This game invites participants to design, build, and optimize LEGO vehicles capable of navigating courses filled with obstacles and challenges. Using scenario cards that simulate race conditions and data biases, players experience the power of Artificial Intelligence algorithms and machine learning by modifying their creations to better meet the course requirements.
​
This game does more than entertain; it educates. By drawing bias cards, participants learn how biases in algorithms can influence outcomes and how to correct them. By sharing their creations on social media, they contribute to a collaborative learning community, enriching their understanding of AI while having fun.
​
Join 'LEGO® Bots: The AI Builders' to build not just vehicles, but also a future where artificial intelligence is demystified and accessible to all. Get ready to launch, test, and refine your AI knowledge, one brick at a time!
01
Objective of the game 'LEGO® Bot
The objective is to build a vehicle capable of rolling and carrying objects, and which moves efficiently on a predefined course, thereby illustrating the construction and outcome of an artificial intelligence algorithm.
02
The Data
The data that the algorithm will process comes in the form of LEGO® bricks of various shapes and functionalities. Participants will receive LEGO® bricks including wheels, axles, and base plates, representing the data that the artificial intelligence will need to assemble and use.
03
Algorithm Programing
Step 1: Integrate a Bias into the Data
Participants use bias cards to introduce a typical challenge in building artificial intelligence solutions that affect the design of their LEGO® vehicle. Each card randomly assigns specific conditions that must be considered during the vehicle construction (small wheels only, only one color...)
Step 2: Data Hunting
This stage introduces participants to the collection and sorting of essential data for building the algorithm. Participants start by gathering LEGO bricks according to a specific list and instructions provided by their bias card, symbolizing the extraction of relevant data. They then organize these 'data' by color, size, or shape to illustrate the data cleaning process.
Step 3: Pseudocode to Structure the Code
In this phase of the game, participants are invited to determine the essential steps for building their vehicle, while minimizing effort. This is known as pseudocode, a simplified version of the steps that will be used. This helps clarify the necessary logic before moving on to programming.
Step 4: Conditional Logic (Triggers and Actions)
Participants receive a set of forecast cards containing challenges to integrate into the design of their LEGO® car models. These conditions act as programming directives that specify how the vehicle should react in given situations.
04
Machine Learning
Participants fine-tune their constructions to optimize efficiency or speed, based on observations of their previous performances. This step replicates the machine learning process by allowing the model to adapt and improve gradually.
05
Open Source sharing
Participants are encouraged to take a photo of their LEGO® AI vehicle as well as the steps of the 'code' of their construction, and to post it on Instagram. They then invite the community to identify any biases that may have been incorporated into the design of their vehicle. This interaction not only promotes the exchange of knowledge and ideas but also stimulates critical thinking about the influences of algorithmic biases.
