Class 9 Ch-4 Introduction to Gen AI PDF

Ch-4 Introduction to Generative AI Learning Objectives To understand Generative AI and its types. To know examples and benefits of using Generative AI. To identify popular Generative AI tools and their applications. To sensitize the students about the ethical considerations of using Generativ...

Ch-4 Introduction to Generative AI Learning Objectives To understand Generative AI and its types. To know examples and benefits of using Generative AI. To identify popular Generative AI tools and their applications. To sensitize the students about the ethical considerations of using Generative AI. To explain students about the potential negative impact of Generative AI on society. Learning Outcomes: Students will be able to define Generative AI & classify different kinds. Students will be able to explain how Generative AI works and recognize how it learns. Students will be able to apply Generative AI tools to create content. Students will understand the ethical considerations of using Generative AI Activity: Guess the Real Image vs. the AI- Generated Image Purpose: Tounderstand the difference between real and AI- Generated Images. Examine the images and determine whether either This is where our analytical skills come into play. We can uncover crucial clues by focusing on the of the images is a real image or an AI generated image. nuances in lighting and texture. If you zoom in, you will notice how simplistic the shuttle design is in the AI-generated image compared to the real image, which features more complexity in the lines, divots, and shading. Also, give reasons for your answer  Three of these images are AI generated and one is an actual photo Supervised Learning and Discriminative Modeling in AI The classification of data elements into categories or labels was initially taught to the machine learning models by humans. Imagine a teacher is training students to identify different types of leaves. The teacher shows pictures of leaves, point out their features (broad, pointy, etc.), and tell the student if it's a mango leaf, neem leaf, and so on. This is similar to how computers learn in supervised learning, a type of Artificial Intelligence (AI). Supervised Learning is like training a computer with examples. We provide the computer with data (pictures of leaves) and tell it the desired outcome (type of leaf). The computer analyzes the data, learns the patterns, and can then predict the outcome for new, unseen data (identifying a new leaf). Here is how it happens: 1. Training Data: We give the computer a lot of data points, like labeled pictures of leaves. Each picture is an instance, and the label (mango leaf, neem leaf) is the target variable. 2. Learning Algorithm: The computer uses a special program to analyze the data. This program is like the student learning from the teacher. 3. Making Predictions: Once trained, the computer can predict the target variable for new, unseen data. It can analyze a new leaf picture and tell you if it's a mango or neem (or something else entirely!). Discriminative Modeling is a specific type of supervised learning where the computer learns to distinguish between different categories. In our leaf example, the computer learns the differences between mango leaves, neem leaves, and other types based on features like shape and size. Here is an image which explains the concept of these images are generating. Source: https://lingarogroup.com/blog/generative-ai-explained-by-humans Examples of Supervised Learning and Discriminative Modeling in Action: Spam Filtering: Your email uses supervised learning to identify spam emails. It analyzes emails and learns the patterns of spam messages (words, sender information) to filter them out. Face Recognition: Social media apps use supervised learning to recognize faces in photos. They are trained on millions of labeled images (person A, person B) to identify people in new pictures. Handwritten Digit Recognition: Banks use supervised learning to read handwritten digits on cheques. The computer is trained on many labeled images of handwritten numbers to recognize them accurately. Activity 1: Sort the Fruit Basket 1. Take pictures of different fruits (apple, orange, banana). 2. Label each picture with the fruit name. 3. Use a free, online image classification tool (available with basic functionality) and upload your pictures. [Scan the QR code to use online image classification tool https://cloud.google.com/vision ] 4. Train the tool to identify the fruits based on your pictures. 5. Test the tool with new pictures - can it correctly identify the fruits? Unsupervised Learning and Generative Modeling Imagine you're at a party with a bunch of masked people. You can't see their faces, but by observing their conversations and interactions, you can guess who might be friends or belong to the same group. This is similar to how computers learn in unsupervised learning, another type of Artificial Intelligence (AI). Unsupervised Learning is like letting a computer explore a big pile of data on its own, without pre- defined categories or labels. The computer finds patterns, similarities, and hidden structures within the data, just like you guessed friendships at the party. This can be further simplified here: 1. Unlabeled Data: We give the computer a lot of data, but this data doesn't have any specific labels or categories. It's like a pile of clothes without tags. 2. Pattern Recognition: The computer uses special algorithms to analyze the data and find patterns or relationships between different data points. 3. Grouping and Clustering: Based on the patterns, the computer might group similar data points together. It's like figuring out who talks to whom the most at the party. Generative Modeling is a specific type of unsupervised learning where the computer uses the patterns it finds to create entirely new data. Going back to the party analogy, after observing interactions, the computer might try to predict what two people might talk about next, or even create a story about the party! Examples of Unsupervised Learning and Generative Modeling in Action: Recommendation Systems: Apps like Netflix or Youtube use unsupervised learning to recommend movies or videos you might like. They analyze your watch history and find patterns to suggest similar content. Image Segmentation: Medical imaging software uses unsupervised learning to identify different tissues or organs in an X-ray or MRI scan. It analyzes the image and groups pixels with similar characteristics. Music Generation: Some AI programs can create new music pieces by learning from existing songs. They analyze the patterns in music (tempo, rhythm, melody) and generate new compositions based on those patterns. Activity 1: Play "I Spy - Unsupervised Edition" 1. One student describes an object in the classroom (chair, table, etc.) without saying its name. 2. The other students ask yes/no questions about the object's properties (e.g., Is it made of wood? Does it have legs?)  3. Based on the questions, they try to guess the object. 4. This is similar to unsupervised learning - they are using the object's properties (data) to identify it (like a computer finding patterns). In unsupervised or self-supervised learning, the machine learning model takes unlabeled datasets and figures out patterns and inherent structures within them — without human intervention. Activity: Watch the video by scanning the QR code and share your views: https://www.youtube.com/watch?v=26fJ_ADteHo _____________________________________________________________________________ _____________________________________________________________________________ _____________________________________________________________________________ What is Generative AI? Generative AI, short for Generative Artificial Intelligence, refers to a type of AI that can create new data, like text, images, code, or even music. Imagine it as a creative AI artist who learns from a massive collection of existing works and then uses that knowledge to produce entirely new and original content. This technology is trained with existing data and content, creating the potential for applications such as natural language processing, computer vision, the metaverse, and speech synthesis. Simply we can say Generative AI is basically a computer program that can create new things, like text, images, or even music, based on what it's learned from existing data. Think of it like a supercharged copycat that can learn styles and patterns from a massive amount of information. What is the foundation for Generative AI? Foundational advancements: The rise of neural networks and deep learning laid the groundwork for the sophisticated generative models we see today. Progressive breakthroughs: From early attempts to recent successes in areas like natural language processing and image generation, generative AI has seen a steady stream of innovations. Continuous refinement: The field is constantly evolving, with researchers improving existing techniques and exploring new applications. Diverse applications: Generative AI's reach now extends far beyond its initial goals, impacting areas like text creation, image synthesis, and even creative content development. Let us have a look at timeline of Generative AI  Source: MI6 website The Evolution of Generative AI: A Journey from Simple to Spectacular Early Steps: Learning from Scratch Back in the day, computers were like robots who only followed orders. But then came neural networks, inspired by the human brain. These networks allowed computers to learn from examples, just like us! This was a key step for Generative AI. Training Makes Perfect: Deep Learning Takes the Wheel Think of learning a new language. The more words you learn, the better you speak it. Similarly, Generative AI needs a lot of data to improve. This is where deep learning comes in. Deep learning allows AI to process massive amounts of information, like text, pictures, and music. By analyzing these "mountains" of data, Generative AI learns patterns and recreates them in new ways. From Simple to Spectacular: Creativity Takes Flight! Remember those old, rule-following computers? With advancements in neural networks and deep learning, Generative AI started getting creative. It could now generate new things based on what it had learned. Imagine feeding it a bunch of poems and then asking it to write its own – that's the power we're talking about! The Rise of the Machines... As Artists? Today, Generative AI can do some pretty cool things: Writing Partner: Stuck on your school project? Generative AI can help craft compelling stories based on your ideas. Creature Creator: Describe a mythical beast, and the AI might create a picture of it coming to life! Music Maestro: Need a catchy tune for your school play? Generative AI can suggest musical pieces based on the mood you describe. The Future is Bright (and Creative!) Generative AI is still under development, but it's already changing the world. It's a powerful tool that sparks your creativity and helps you explore new ideas. Remember, the possibilities are endless! So, the next time you're feeling uninspired, consider giving Generative AI a try. You might be surprised at what you create together! Explore: What do you understand about generative AI? _____________________________________________________________________________________ _____________________________________________________________________________________ Give a few examples of generative AI. _____________________________________________________________________________________ _____________________________________________________________________________________ What do you know about Deep Fake? _____________________________________________________________________________________ _____________________________________________________________________________________ Generative AI vs. Conventional AI: A Tale of Two Intelligences Unlike conventional AI, which excels at analyzing and classifying existing data, Generative AI is a game- changer. It's specifically designed to create entirely new and unique content, pushing the boundaries of creativity. Here's how they differ: Generative AI focuses on generating fresh content, from realistic portraits of GOAL non-existent people to catchy tunes. Conventional AI, on the other hand, is the master of analysis, processing, and categorizing existing data. Generative AI models are fueled by massive datasets. These datasets train TRAINING complex structures like neural networks to identify patterns and generate new content based on those learnings. Conventional AI relies on less complex algorithms and training methods focused on specific tasks. The magic of Generative AI lies in its ability to produce unexpected and OUTPUT innovative content. Think of it as an artist constantly exploring new possibilities. Conventional AI, however, delivers predictable outputs based on the data it has been trained on. Generative AI is transforming creative fields like art, music, literature, and APPLICATION design. It can even generate realistic video game environments! In contrast, conventional AI finds its applications in areas like banking, healthcare, image recognition, and language processing. Types of Generative AI Generative AI is a type of artificial intelligence that can create new data. Here are some common types of generative AI: 1. GANs (Generative Adversarial Networks) How they work:  GANs are neural networks that collaborate to produce fresh data Made up of two neural networks: A Generator Network and a Discriminator Network. The generator creates new data, while the discriminator checks if the data looks real or fake. They keep improving until the generator can create data that looks just like real data. Examples: Creating pictures of people who don't exist. Turning daytime photos into night time photos. Making images based on text descriptions. Creating realistic video scenes. 2. VAEs (Variational Autoencoders) AI Generated Image How they work: VAEs learn the patterns in data and use them to create new, similar data. They compress data into a smaller format and then expand it back to its original form. Examples: Creating new images that look like the ones in the training set. Reconstructing damaged images. Generating draft text for writers. Creating new sounds and music. 3. RNNs (Recurrent Neural Networks) How they work: RNNs are designed to handle sequences of data, such as text or music. They remember past inputs and use that information to predict future inputs. Examples: Writing new text in the style of a famous author. Predicting the next word or character in a sentence. Composing music. 4. Autoencoders How they work: Autoencoders learn to compress data and then decompress it to its original form. They can remove noise from images or reduce the size of image files. Examples:  Creating artistic images. Discovering new medicines. Unleashing Creativity: The Power of Generative AI Generative AI is transforming the way we interact with the world, injecting a dose of artificial ingenuity into creative fields. From art exhibitions to music concerts, its influence is undeniable. Art Reborn: The Next Rembrandt Imagine a new Rembrandt painting – not a hidden masterpiece unearthed from an attic, but a meticulously crafted creation born from artificial intelligence. The Next Rembrandt project is a testament to this possibility. By analyzing vast amounts of data from Rembrandt's existing works, researchers were able to train a generative AI model to produce a painting that captures the essence of the artist's style. The result? A brand new piece that could seamlessly blend into Rembrandt's collection. (Scan the QR code to watch video: Video source: The Next Rembrandt. (2016, April 5). YouTube. https://www.youtube.com/watch?v=IuygOYZ1Ngo ) AI Symphony: AIVA's Musical Compositions Music lovers, prepare to be surprised. Generative AI is not just mimicking the old masters; it's composing entirely new pieces. AIVA, an AI composer, takes the stage with its ability to craft original music in various genres. From melancholic piano sonatas to upbeat electronic dance tracks, AIVA demonstrates the vast potential of AI to create music that is both innovative and evocative. (Scan the QR code to watch video: Video source: TED. (2018, October 1). How AI could compose a personalized soundtrack to your life | Pierre Barreau. YouTube. https://www.youtube.com/watch?v=wYb3Wimn01s ) Language Unleashed: Chatbots and Natural Language Generation The way we communicate is also undergoing an AI-powered transformation. Generative AI is behind the chatbots that can hold conversations with us, answering our questions and fulfilling our requests. It's also fueling natural language generation systems, capable of producing human-quality written content. These are just a few examples of how generative AI is pushing the boundaries of creativity. As this technology continues to evolve, we can expect even more remarkable applications to emerge, blurring the lines between human ingenuity and machine intelligence. (Watch video: Video source: BBC News. (2023, January 15). What is ChatGPT, the AI software taking the internet by storm? - BBC News. YouTube. https://www.youtube.com/watch?v=BWCCPy7Rg-s ). Benefits of Generative AI Generative AI isn't just about creating cool art or music; it's a game-changer across industries. Here's why: Boosting Creativity: Stuck on a design concept? Generative AI can spark new ideas and help artists, designers, and musicians break through creative roadblocks. Efficiency on Autopilot: Repetitive content creation tasks? Let AI handle them. Generative AI can automate processes, saving time and resources for more strategic endeavors.  Personalization Powerhouse: Imagine content tailored to every user. Generative AI can create personalized experiences, from product recommendations to news feeds, keeping users engaged. Exploration Unleashed: Need to explore a million design possibilities? Generative AI can analyze data and suggest innovative solutions, optimizing complex systems like drug discovery or industrial processes. Accessibility for All: Generative AI empowers everyone. It puts sophisticated content creation tools within reach, allowing anyone to produce high-quality content regardless of expertise. Scalability Made Simple: Need tons of content, fast? Generative AI pumps out large volumes efficiently, making it perfect for businesses with high content demands. From sparking creativity to automating tasks, generative AI offers a multitude of benefits. As this technology matures, we can expect even greater transformations across various fields. Limitations of Using Generative AI Generative AI is a revolutionary tool, but like any powerful technology, it has limitations to consider: Data Bias or Data Dilemmas: Generative AI is only as good as the information it's trained on. Biased or incomplete data sets can lead to biased or inaccurate outputs. Imagine a facial recognition system trained on a limited dataset – it might struggle to recognize faces of certain ethnicities. The Unpredictable Muse(Uncertainty): While generative AI can spark creativity, its outputs can be unpredictable. Sometimes it might surprise you with a novel idea, but other times it might generate nonsensical content. Computational Demands: Training and running generative AI models requires serious computing muscle. This can be expensive and time-consuming, especially for smaller businesses or individual users. Generative AI requires significant computational resources to train and generate its output. Hands-on Activity: GAN Paint Ready to experiment with generative AI yourself? Check out GAN Paint! Select a Base Image: Choose a starting point from the website's library. Brush it Up: Use the brush tool to add objects and textures to your image. AI Learns, You Create: As you paint, the GAN network will learn to generate more realistic images based on your input. Explore GAN Paint by scanning the QR code: https://ganpaint-v2.vizhub.ai/ and see what creative visions you can bring to life! Generative AI and Its Tools Generative AI is transforming creation, and there are a wealth of tools available to explore its potential. Let's dive into some popular options: Artbreeder: The Art Mashup Studio Artbreeder is a web-based playground for generating new images by combining different AI models. Imagine selecting traits from various artistic styles and merging them – a cubist portrait with pop-art colors? Artbreeder makes it possible! Hands-on Activity: Generate with Text Prompts 1. Head to Artbreeder. 2. Click "Create" and choose "New Image" under "Prompter." 3. Unleash your creativity with a cool text prompt – what image comes to mind? See how the AI translates your words into a picture! Runway ML: Building Your Own AI Magic Runway ML takes things a step further, offering a platform to create, train, and deploy generative models. It empowers users with a user-friendly interface to build and train various AI models, like those for generating images or classifying objects. Explore the AI Magic Tools 1. Visit Runway ML by scanning the QR Code. 2. Dive into the "AI Magic Tools" section. 3. Pick a tool that sparks your interest and experiment with generating new content! ChatGPT: Chatting with AI (Link: https://chat.openai.com/) ChatGPT is an example of a large language model, a type of generative AI that can produce human-quality text. Here's an example of how it introduced itself when I queried it: I am a large language model chatbot developed by OpenAI. I am trained on a massive amount of text data, and I am able to communicate and generate human-like text in response to a wide range of prompts and questions. For example, I can provide summaries of factual topics or create stories. ▪ Gemini Link: https://gemini.google.com/ Scan the QR code to visit Gemini AI I asked Gemini to introduce itself. And here is the response! Hands-on Activity Chit-Chat with ChatGPT & Gemini ▪ Sign up & Login into both ChatGPT and Gemini. ▪ Chat with the ChatGPT and ask it to write a paragraph on How it Works? - ChatGPT ▪ Similarly, Chat with Bard and ask it to write a paragraph on How it Works? - Gemini Here are 6 prompts that can be tried on ChatGPT and Gemini: 1. Write a summary of the history of the internet. 2. Explain how to code a simple website. 3. Write a blog post about the latest trends in artificial intelligence. 4. Create a presentation about the benefits of cloud computing. 5. Write a research paper about the future of technology. 6. Design an app that solves a real-world problem. Document the findings from above activity on ChatGPT vs Gemini vs Copilot based on the parameters below: ▪ Parameter 1: Human-Like Response. ▪ Parameter 2: Training Dataset and Underlying Technology. ▪ Parameter 3: Authenticity of Response. ▪ Parameter 4: Access to the Internet. ▪ Parameter 5: User Friendliness and Interface. ▪ Parameter 6: Text Processing: Summarization, Paragraph Writing, Etc. ▪ Parameter 7: Charges and Price. How to Use Generative AI Tools in Real-world Scenarios. The table shows popular Generative AI tools that can be used in various fields. Generative AI: Power and Responsibility Generative AI offers a treasure trove of possibilities, but like any powerful tool, it requires careful consideration. Let's delve into the ethical considerations surrounding its use: Potential Societal Impact Misinformation Warfare: Generative AI can be misused to create fake news or deepfakes, manipulating public opinion and eroding trust. Job Displacement Concerns: As AI automates content creation tasks, job displacement for humans in these fields becomes a concern. Data Security Risks: Generative AI could potentially be used to create sensitive personal information, posing a privacy and security threat. Ownership and Authorship: Who Owns the Muse?: As generative AI creates original works, questions arise about ownership, particularly in creative fields like art or music. Where does human authorship end and machine authorship begin? Human Agency and Control Blurred Lines of Origin: As AI sophistication advances, distinguishing human-generated content from machine-generated content becomes increasingly difficult. This could lead to a loss of human control and autonomy. Bias and Fairness Amplifying Bias: Generative AI models trained on biased data can perpetuate those biases in their outputs. This can have harmful consequences, especially in high-stakes applications like loan approvals or criminal justice. Combating Misinformation Fake News Factories: Generative AI can be weaponized to create fake news and manipulate public opinion. This undermines trust in institutions and democratic processes. Privacy Concerns Personal Information at Risk: Generative AI has the potential to generate sensitive personal data, like credit card numbers, raising privacy and security concerns. The Path to Responsible Use Diverse Training Data: Ensure the data used to train generative AI models is diverse and representative to avoid perpetuating biases. Scrutiny and Fact-Checking: Scrutinize outputs for bias and misinformation to prevent their spread. Prioritizing Privacy: Prioritize user privacy and obtain informed consent when using generative AI. Clear Ownership Guidelines: Establish clear guidelines around ownership and attribution of content created with generative AI. Open Public Dialogue: Engage in public discussions about the social and ethical implications of generative AI to guide its development for societal benefit. By emphasizing ethics, fostering trust, minimizing negative impacts, establishing clear regulations, and encouraging responsible innovation, we can harness the immense potential of generative AI to create a better future. Exercise A. Multiple Choice Questions (MCQs) 1. Generative AI can create new data in the form of: a) Text only b) Images only c) Text, Images, Code, and Music (Correct) d) Numbers only 2. Generative AI models are trained on: a) Existing user data b) Instructions c) Existing data (Correct) d) Random patterns 3. Generative Adversarial Networks (GANs) are a type of: a) Search Engine b) Generative AI model (Correct) c) Data Compression tool d) Social Media Platform 4. Which of the following is NOT a benefit of Generative AI? a) Increased Efficiency b) Boosted Creativity c) Data Bias (Correct) d) Personalized Experiences 5. Generative AI is limited by:  a) User Interface complexity b) High Availability c) Computational Demands (Correct) d) Limited internet access 6. Popular Generative AI tools include: a) Artbreeder b) Runway ML c) ChatGPT (Correct) d) All of the above 7. Generative AI can be used to: a) Improve medical diagnosis (Correct) b) Write fake news articles c) Neither a nor b d) Both a and b 8. Generative AI models work by: a) Following a set of pre-programmed rules b) Identifying patterns in data (Correct) c) Randomly generating content d) Connecting to the internet B. Fill in the Blanks 1. Generative AI is a type of artificial intelligence that can create new ________. (Answer: Data) 2. The process of training a generative AI model involves feeding it a large amount of ________. (Answer: Data) 3. Generative AI models can be used to ________ creativity. (Answer: Boost) 4. One of the limitations of generative AI is ________ bias in the training data. (Answer: Data) 5. Popular generative AI tools include ________ for creating artistic images. (Answer: Artbreeder) 6. Generative AI has the potential to ________ many industries. (Answer: Revolutionize) 7. A generative AI model called a _____________ can generate realistic images. (Answer: Generative Adversarial Network (GAN)) 8. Generative AI is a powerful tool, but it is important to use it __________. (Answer: Responsibly) C. Short Answer Questions 1. Briefly explain how generative AI models work. Answer: Generative AI models are trained on a large amount of data. They learn to identify patterns in the data and then use these patterns to generate new, original content. 2. What are some of the benefits of generative AI? Answer: Benefits include boosting creativity, improving efficiency, personalizing experiences, and exploring new possibilities. 3. What are some of the limitations of generative AI? Answer: Limitations include data bias, unpredictable outputs, and high computational demands. 4. Give two examples of how generative AI can be used in real-world scenarios. Answer: Examples include creating personalized learning materials, generating new product designs, or composing music. 5. Why is it important to be aware of the limitations of generative AI?  Answer: By understanding the limitations, we can use generative AI more effectively and responsibly, avoiding potential biases or unintended consequences. D. Long Answer Questions 1. Describe the different types of generative AI models and explain how they differ from each other. Answer: Hint: Discuss Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and Recurrent Neural Networks (RNNs), highlighting their strengths and weaknesses. 2. Discuss the ethical considerations surrounding the use of generative AI. How can we ensure that generative AI is used responsibly? (Answer: Consider issues like data bias, potential for misuse, and the impact on creative industries. Discuss strategies for responsible use such as data transparency, bias mitigation techniques, and clear guidelines for developers and users. 3. Imagine you are a product manager tasked with developing a new generative AI tool. What specific problem would your tool address, and how would it use generative AI technology to solve it? Answer: [Hint] Describe the problem your tool addresses (e.g., generating personalized music recommendations, creating realistic simulations for training purposes E. Assertion and Reasoning Instructions: Evaluate each statement and choose the most appropriate answer. 1. Assertion (A): Generative AI can be used to create realistic fake news articles. Reason (R): Generative AI models can be trained on large amounts of text data. a) Both A and R are true, and R is a correct explanation of A. b) Both A and R are true, but R is not a correct explanation of A. (Correct) c) A is true, and R is false. d) A is false, and R is true. 2. Assertion (A): Generative AI is a threat to creative jobs. Reason (R): Generative AI can create content like music and art. a) Both A and R are true, and R is a correct explanation of A. b) Both A and R are true, but R is not a correct explanation of A. (This is a possibility, but generative AI can also assist creative professionals.) c) A is true, and R is false. d) A is false, and R is true.  3. Assertion (A): Generative AI models require very little computational power to run. Reason (R): Generative AI models are becoming increasingly complex. a) Both A and R are true, and R is a correct explanation of A. b) Both A and R are true, but R is not a correct explanation of A. (This contradicts R.) c) A is true, and R is false. (This contradicts R.) d) A is false, and R is true. (Correct) F. Case Study A fashion company is looking to use generative AI to improve its design process. They want to create a tool that can generate new clothing designs based on current trends and customer preferences. Questions: 1. What type of generative AI model would be most suitable for this task? Explain your answer. 2. How would the fashion company go about training the generative AI model? What kind of data would they need to collect? 3. What are some potential benefits and drawbacks of using generative AI in the fashion design process?** Answer: 1. A suitable generative AI model for this task could be a Variational Autoencoder (VAE). VAEs are good at learning latent representations of data, which can be useful for generating new variations based on existing styles. 2. To train the model, the fashion company would need to collect a large dataset of existing clothing designs. This data could include images of clothing items, along with information on styles, materials, colors, and customer preferences (e.g., through surveys or sales data). 3. Benefits: o Generate a wider range of design options for exploration. o Improve design efficiency by automating some tasks. o Identify and capitalize on emerging trends based on data analysis. Drawbacks: o Reliance on training data quality – biased data can lead to biased designs. o The “creative spark” might be lost if designers rely solely on AI-generated suggestions. o The need for expertise in managing and interpreting AI outputs.