AI Innovation Corner
These articles are part of the FIRST Tech Challenge AI Innovation Corner. This is a place where we’ll post custom and curated articles relevant to FIRST Tech Challenge as it relates to AI and its impact on our daily lives and the world around us. We would like to thank Google for their generous contributions to FIRST Tech Challenge to increase access to our program in underserved communities and for providing sponsorship and occasional technical direction for this content.
Articles are ordered on this page chronologically, with the newest content at the top of the page expanded. Just click to expand any other articles you’d like to see.
Week of 09/30/2024 “AI Competition Manual Assistant”
AI Competition Manual Assistant
In our first article, the Google AI Studio was introduced as a tool to interact with Google’s Gemini AI. Gemini is one of several flagship Large Language Models (LLM’s) that have been meticulously trained on massive amounts of text data to learn the patterns and relationships between units of language - these models have actually learned how to recognize text-based language, read and understand data, and synthesize what it learned to predict and interpret future data. This is the exact process humans make in learning and understanding the world around us! In Google AI Studio, users can interact with the Gemini AI through “prompts” to perform tasks for them. Prompts are instructions or queries given to an AI in order to generate a response - the quality of the response is often directly related to the quality of the prompt. Through these prompts, Gemini can provide responses based on the massive dataset that it has been pre-trained with, or users can also provide additional documents, text, or media that the AI has never seen before. These multimodal prompts, or prompts that include multiple types of content, can be very beneficial in interacting with an AI using content that is specific to a niche area like FIRST Tech Challenge. Can you think of ways to put this ability to good use in FIRST Tech Challenge?
In FIRST Tech Challenge, one of the first tasks teams have to do is to read and understand the FIRST Tech Challenge Competition Manual. This can be a very painstaking task, and even a skilled reader can miss subtle nuances provided by the manual. However, an AI can break down and analyze the manual in a matter of seconds, usually preserving the nuance provided in the document. Users can then interact with the AI that has analyzed the Competition Manual, and prompt the AI to provide insights - these questions might involve locating specific information likely found in the Competition Manual, summarize important rules or processes, or even involve asking the AI to make a best guess. Through a process known as “role playing” the user can prompt the AI to take on a role or persona and direct the AI to follow specific rules as it interacts with the user in subsequent prompts. The remainder of this article is a tutorial on how to set up a “role playing” session with the Google Gemini AI through Google AI Studio to analyze and answer questions based on the FIRST Tech Challenge 2024-2025 Competition Manual for the INTO THE DEEP presented by RTX season. While some of the nuanced elements (like AI prompting) will be shallowly covered in this article, it is something we’ll cover a lot more in future articles.
Creating an AI expert using Google AI Studio is fairly straightforward - the hard part is creating the proper prompt, and there we’ve got you covered.
Step 1 - First, log into Google AI Studio. You can do this by clicking the “Sign in to Google AI Studio” button on the front page of the Google AI Studio home page. You will need a Google account in order to do this - getting one is left as an exercise to the reader. The Google account is used to store your Google AI Studio prompt sessions and any content you upload to the model, and to track usage of the Gemini APIs.
Step 2 - Let’s download the FIRST Tech Challenge Competition Manual to your local computer. You can always find the latest Competition Manual PDF at the following link:
Step 3 - In the left navigation pane towards the top of the pane, there is a circle with a plus inside it with the text “Create new prompt” next to it. Clicking on this button will start a new prompt - though if you’re using Google AI Studio for the first time it’s likely a new prompt is already open.
Now that we have a new prompt, you can give the prompt a name. This will allow the prompt to be saved in your “My Library” so you can come back and interact with the prompt later without having to recreate the prompt session every time.
In the bottom center of the workspace is a text field where you can enter in your prompt (it has a default prompt of “Type something”). BEFORE we enter our prompt, we want to add our Competition Manual PDF document. To add the document, click the “Plus” icon to the right of the prompt area. This will give you several options, choose “Upload to Drive”. You can either click the “Browse” button to browse for the PDF of the Competition Manual that you downloaded, or you can drag the file into the window. This adds the Competition Manual to your prompt, it may take a minute or two to upload the PDF so please be patient.
Step 4 - Now that we have our document uploaded, we now want to enter our prompt. This prompt directs the AI in how to manage its responses, what information to use when developing a response, and sets up the role that the AI will attempt to play. Enter the following prompt and press the “Run” button:
You are a helpful AI assistant providing answers to questions about the provided PDF. Do not use any prior knowledge; you have everything you need to answer questions in the one PDF provided. Cite all references.
Once the AI processes the initial prompt, we can then ask questions that the AI will use the Competition Manual to answer. Depending on the question, it may take the AI between several seconds up to a couple minutes to answer - be patient! Here are several questions you can ask (remember to press the “Run” button after asking each question):
Example sample questions:
How many SAMPLES is a ROBOT allowed to CONTROL at a time?
What are the different ways to score points?
How large can a ROBOT be in its STARTING CONFIGURATION?
Which awards are best for advancement?
How do I write a strong engineering portfolio?
Some prompts that require a lot of complex understanding or strategy can yield results that are not correct, especially if there is information “understood but not supplied.” For example, the following prompts provide some correct and some incorrect information:
Examples of difficult questions:
What is the maximum score for an alliance?
Can ROBOTS pick up an opposing ALLIANCE’S SAMPLES?
How many matches does a team play at an event?
This example was specific to FIRST Tech Challenge, but this process can be used for virtually any documents or media. Using AI as an analysis assistant can help you summarize news articles, find specific instructions in user manuals, review books, and more! Remember that the quality of the responses the AI provides is directly related to the quality of the prompt provided - even so, the AI isn’t always going to be able to provide correct answers so it’s up to you to verify the correctness of all answers provided by an AI.
Week of 09/09/2024 “AI Innovation Corner - Google AI Studio”
AI Innovation Corner - Google AI Studio
This first article launched as part of the Tech Tips of the Week, but is the official first article for the AI Innovation Corner.
This week’s Tech Tip of the Week launches a new initiative in FIRST Tech Challenge, an AI Innovation Corner. Generative AI has taken the world by storm, becoming commonplace now in everything from personal assistants, search engines, recipe curation, music innovation, and vehicle maintenance! Machine Learning AI has been a part of FIRST Tech Challenge in some way for the past six years, and we’re now transitioning to help teams learn how to use and incorporate Generative AI in their FIRST Tech Challenge experience (while we’re learning ourselves!).
The first step (or FIRST step?) to getting the most out of AI is choosing a model. What do I mean by model? Every AI is a neural network that has been trained with specific knowledge with the ability to do specific things based on that knowledge. Each version of this neural network is stored in a “model”. Each different company has different models available for different purposes, though most models are variations on their flagship model (Gemini from Google, ChatGPT 4-o from OpenAI, Claude from Anthropic, and so on). Each company has different web-based and API interfaces for interacting with their models, and everyone has their favorite. In FIRST Tech Challenge, the standard tool we use is Google AI Studio to interact with Gemini.
Google AI Studio is free to use, but requires a Google account to access - virtually all models require a login or API token of some kind to use. Google AI Studio is our favorite for its list of examples (Prompt Gallery) and its easy to use interface to save prompt sessions and resume them later. With Google AI Studio, you also can select the specific model you want to use, and when available you can choose to use preview versions of up and coming models.