Build Your First AI Model: A No-Code Teachable Machine Intro

Discovering AI is your guide to understanding artificial intelligence and how it’s rapidly changing the technology we use every day. From suggesting what to watch next to helping doctors diagnose diseases, AI is already deeply woven into the fabric of our lives. But how does it actually work? Are we really on the cusp of a robot takeover, or is the reality far more nuanced (and frankly, less exciting)? For a deeper dive, read AI Unveiled: How It Works.

1. Setting Up Your AI Playground: Teachable Machine

Want to get your hands dirty right away? Then let’s start with Google’s Teachable Machine platform. It’s a free, browser-based tool that lets you train simple AI models without writing any code. We’re going to build an image recognition model that can tell the difference between a picture of a dog and a picture of a cat. Yes, it’s basic, but it’s a great way to grasp the fundamental concepts.

1. Go to the Teachable Machine website.
2. Click “Get Started.”
3. Choose “Image Project.”
4. Select “Standard Image Model.”

You’ll see three sections labeled “Class 1,” “Class 2,” and “Add a class.” Each class represents a category your AI will learn to identify.

5. Rename “Class 1” to “Dog.”
6. Rename “Class 2” to “Cat.”

Now, you need to feed the machine some data. For each class, you can either upload images from your computer or use your webcam to take pictures directly. The more images, the better the model will perform.

7. Click “Upload” under the “Dog” class and select at least 50 pictures of dogs.
8. Click “Upload” under the “Cat” class and select at least 50 pictures of cats.

Pro Tip: Vary the images you upload. Use different breeds, angles, lighting conditions, and backgrounds. The more diverse your training data, the more accurate your model will be.

Once you’ve uploaded your images, it’s time to train the model.

9. Click the “Train Model” button.

The training process can take a few minutes, depending on the size of your dataset. Once it’s done, you can test your model using the “Preview” section. Upload a new image of a dog or cat, or use your webcam. The model will display its prediction, along with a confidence score.

Common Mistake: Not uploading enough images. If your model is consistently misclassifying images, it probably needs more training data. Aim for at least 100 images per class for better results.

2. Talking to AI: Exploring Pre-Trained Language Models

Image recognition is cool, but what about natural language processing (NLP)? Fortunately, you don’t need to build your own language model from scratch. There are many pre-trained models available that you can use for various tasks, such as text generation, translation, and sentiment analysis. One of the most popular libraries for working with these models is Hugging Face’s Transformers library.

First, you’ll need to install the Transformers library. Open your terminal or command prompt and run the following command:

pip install transformers

Once the installation is complete, you can start using pre-trained models. Let’s try generating some text using the GPT-2 model.

1. Open a Python interpreter or create a new Python script.
2. Import the `pipeline` function from the `transformers` library:

from transformers import pipeline

3. Create a text generation pipeline using the GPT-2 model:

generator = pipeline('text-generation', model='gpt2')

4. Generate some text by providing a starting prompt:

prompt = "The quick brown fox jumps over the lazy"
generated_text = generator(prompt, max_length=50, num_return_sequences=1)

In this example, we’re asking the GPT-2 model to generate text starting with the prompt “The quick brown fox jumps over the lazy.” The `max_length` parameter specifies the maximum length of the generated text, and the `num_return_sequences` parameter specifies the number of sequences to generate.

5. Print the generated text:

print(generated_text[0]['generated_text'])

Run the script, and you’ll see the GPT-2 model generate text based on your prompt. The results can be surprisingly coherent (and sometimes hilariously nonsensical).

Pro Tip: Experiment with different prompts and model parameters to see how they affect the generated text. Try using different pre-trained models, such as BERT or RoBERTa, for other NLP tasks. You might also want to read NLP Beyond Chatbots to learn more.

3. Understanding the Ethical Minefield of AI

AI isn’t just about cool technology; it also raises serious ethical questions. One of the biggest concerns is bias. AI models are trained on data, and if that data reflects existing biases in society, the model will likely perpetuate those biases. For example, facial recognition systems have been shown to be less accurate for people of color, leading to potential misidentification and discrimination. A 2023 study by the National Institute of Standards and Technology NIST found significant disparities in the accuracy of facial recognition algorithms across different demographic groups.

Here’s what nobody tells you: AI bias isn’t always malicious. It often arises unintentionally from biased data or flawed algorithms. But the consequences can still be harmful. We had a client last year, a recruiting firm in Buckhead, that implemented an AI-powered resume screening tool. They thought it would save time and improve efficiency. What they didn’t realize was that the model was trained on historical hiring data that favored male candidates. As a result, the AI system was unfairly filtering out qualified female applicants. It took a lawsuit and a lot of bad press before they finally addressed the problem.

Another ethical concern is the impact of AI on jobs. As AI becomes more capable, it’s likely to automate many tasks currently performed by humans. This could lead to job displacement and increased economic inequality. According to a 2025 report by the Brookings Institution Brookings, AI could automate up to 25% of jobs in the Atlanta metropolitan area by 2030. Of course, AI will also create new jobs, but there’s no guarantee that those jobs will be accessible to everyone. The challenge is to ensure that the benefits of AI are shared broadly and that workers are given the training and support they need to adapt to the changing job market.

Common Mistake: Assuming that AI is objective and unbiased. AI models are only as good as the data they’re trained on. It’s crucial to critically evaluate the data and algorithms used in AI systems and to be aware of potential biases.

4. Navigating the Legal Landscape of AI in Georgia

The legal framework around AI is still evolving, but some key areas are already taking shape. In Georgia, for instance, there’s growing interest in regulating the use of AI in specific sectors, such as healthcare and criminal justice. O.C.G.A. Section 51-1-56, the Georgia Tort Claims Act, could potentially be used to hold AI developers liable for damages caused by their systems, although this is still a developing area of law.

Data privacy is another crucial aspect. The Georgia Information Security Act of 2018 (O.C.G.A. § 10-1-910 et seq.) imposes requirements on businesses that collect and store personal information. As AI systems increasingly rely on vast amounts of data, compliance with these laws becomes even more important. We ran into this exact issue at my previous firm. We were advising a startup that was developing an AI-powered marketing platform. They were collecting data on consumer behavior without fully understanding the privacy implications. We had to help them revise their data collection practices to ensure compliance with Georgia law and avoid potential lawsuits.

Intellectual property rights are also relevant. Who owns the output generated by an AI system? Is it the developer of the AI model, the user who provided the input, or someone else entirely? These questions are still being debated, and the answers will likely depend on the specific circumstances. The Fulton County Superior Court has seen a few cases involving AI-generated content, but the legal precedents are still unclear.

Pro Tip: Stay informed about the latest legal developments in AI. Consult with an attorney to ensure that your AI systems comply with all applicable laws and regulations. For more on future trends, see our article AI in 2026: Promise vs. Peril for Business.

5. Staying Informed: Resources for Continued Learning

The field of AI is constantly evolving, so it’s essential to stay informed about the latest developments. Here are some resources that can help you continue your learning journey:

• AI Safety Research Center: AI Safety Research Center This organization focuses on researching and mitigating the potential risks of advanced AI systems.
• Hugging Face: Hugging Face A community and platform dedicated to natural language processing and machine learning.
• MIT Technology Review: A reputable publication that covers emerging technologies, including AI.
• Association for the Advancement of Artificial Intelligence (AAAI): A professional society dedicated to promoting research in AI.

Discovering AI is your guide to understanding artificial intelligence and its potential is a journey, not a destination. By experimenting with tools like Teachable Machine and exploring the ethical implications, you can gain a solid foundation in this transformative technology. Take the leap and start building your own AI projects today – even a simple one can open your eyes to the possibilities.