At Prime Minister, a school of politics for young women, I had the privilege to introduce a group of young, aspiring women to the nuts and bolts of artificial intelligence.
Our session began by unraveling the capabilities of TensorFlow Playground, an interactive web tool that demystifies the workings of neural networks. Together, we selected datasets, adjusted parameters, and observed how changes in the number of layers, neurons, and other variables influenced the model's ability to classify data. It was a digital sandbox where we could forge a tangible connection with the abstract concepts of machine learning.
Next, we simulated Reinforcement Learning from Human Feedback using an approachable proxy: conversing with ChatGPT. Through structured prompts and guided interactions, the students experimented with teaching the model to respond more accurately and ethically. They played the role of both teacher and student, understanding the delicate balance of guiding an AI towards desired behaviors.
Finally, we engaged with the cartography of generative AI, mapping out the terrain of this field's societal impacts. This critical analysis equipped these future leaders with a multifaceted understanding of AI's footprint in society.
Our goal was to empower these young minds with knowledge and practical skills to navigate and shape the AI-driven world. By the end of our time together, these young women were more informed about AI and equipped to assess its applications and implications critically.