´ó·¢

Generative AI

What Is Generative AI? (GenAI)

Generative artificial intelligence (GenAI) is a branch of AI that can create new content or data, such as images, text, music, or speech. Examples of this technology include , , , and .  

One of the main techniques in generative AI is to use large language models (LLM), which are computer programs that can learn from a huge amount of text data and generate new text, images, music, and other content. For instance, Microsoft Copilot and ChatGPT both utilize an LLM called . A large language model is trained by analyzing millions or billions of words from books, articles, websites, or other sources, and learning the patterns and rules of natural language. Then, it can produce new content by predicting the most likely words or sentences that follow a given input, such as a word, a phrase, or a question. Researchers are still working on understanding how these tools work and why they behave the way that they do. 

Potential Benefits

Generative AI has many potential applications such as: 

• Creating realistic and diverse texts for professional or creative purposes (essays, reports, fiction, poetry, etc.) 
• Answering questions, classifying and explaining concepts, or providing information on any topic or domain. 
• Summarizing or paraphrasing long or complex texts into shorter or simpler ones. 
• Translating texts from one language to another, or from one style to another.
• Creating images of videos from text prompts

Advanced Utilization

Generative AI can also be used in more complex endeavors, such as:

• Enabling educators and students to create personalized and interactive learning materials, such as quizzes, flashcards, or simulations.
• Helping researchers and engineers to generate new data for testing or improving their models, such as synthetic faces, voices, or medical images. 
• Creating "autonomous" AI tools (often called "agents") to perform automated functions. 

Challenges & Risks

Generative AI also poses some challenges and risks, such as: 

• Being used for malicious purposes, such as generating fake news, deepfakes, or spam. 
• Raising ethical and social issues, such as copyright, privacy, bias, or authenticity. 
• Affecting human creativity and agency, such as by replacing or influencing human decisions, preferences, or emotions. 
• Increasing energy consumption due to the demands of data centers required to train LLMs.