2. AI Fundamentals - Differentiate between the types of AI and their capabilities.

• Generative AI and predictive AI are distinct types of AI, each with unique capabilities.
• Generative AI creates new content like images, text, video, and code, finding applications in creative fields.
• Predictive AI uses large datasets to identify patterns and forecast future outcomes, suitable for finance, healthcare, and marketing.
• Both types rely on machine learning, but serve different purposes.

Generative AI

• Aims to generate new, original content.
• Trained on diverse and comprehensive data.
• Learns patterns and relationships in data.
• May lack specificity or struggle to filter irrelevant data.
• Exhibits high creativity in content generation.
• Employs complex algorithms and deep learning.
• Responds to prompts by generating images, text, video, and software code.
• Uses prompts and training data to create original outputs.
• Key application: Content generation (chatbots, art, code).
• Employs machine learning systems, algorithms, and neural networks.
• Trained on large datasets to analyze patterns and relationships, enabling new content generation.
• Can analyze literary works and replicate an author's style.
• Enhances data augmentation, personalization, and accessibility.
• Can be misled by crafted input data, leading to incorrect outputs.
• Struggles with lengthy passages, leading to comprehension issues and inconsistent responses.
• Can replicate biases present in training data, affecting content quality.
• Use cases rely on content creation such as text, images, video, audio, data augmentation and personalization

Generative Adversarial Networks (GANs)

• Deep learning framework used in generative AI.
• Consists of a generator and a discriminator, trained competitively.
• The generator creates output, and the discriminator evaluates its authenticity.
• Feedback from the discriminator improves the generator's content generation ability.
• Employs machine learning inspired by past creativity, potentially raising copyright concerns.

Predictive AI

• Aims to predict future outcomes or trends.
• Uses historical data for prediction.
• Learns from historical data.
• Limited to existing patterns and may ignore alternative scenarios.
• Lacks content creation capabilities.
• Relies on statistical algorithms.
• Analyzes past data to make inferences and predict future outcomes and trends.
• Relies on extensive historical datasets to spot patterns over time.
• Key application: Forecasting (sales trends, customer behavior, risk assessment).
• Models utilize historical data, patterns, and trends to predict future events.
• Model development involves data collection, preprocessing, and division into training and testing sets.
• Training involves machine learning algorithms that are iteratively adjusted.
• Accuracy depends on the quality and quantity of training data.
• Can predict future trends, opportunities, and threats.
• Can predict trends, recommend/upsell products, and improve customer service.
• Adds depth and accuracy to management processes.
• Allows businesses to extract more value from data.
• Heavily relies on the availability and quality of data.
• Incomplete, inaccurate, or biased training data leads to flawed predictions.
• Ethical concerns arise regarding privacy, bias, and discrimination.
• Models can be complex, making it challenging to understand predictions or identify biases.
• May not be practical or feasible for all businesses due to computational costs.
• Use cases rely on historical trends and prediction such as financial forecasting, healthcare predictions and marketing strategies

Future of AI

• The distinction between generative and predictive AI is fading.
• Integration of both types enhances business solutions.
• Combining trend identification with creative content generation is a growing trend.

AI Capabilities

• Pattern recognition identifies regularities in data.
• Predictions forecast future outcomes based on historical data.
• Recommendation systems provide personalized recommendations.
• Classifications identify the nature of data, such as phishing emails or fraud.
• Robotic navigation involves real-time adaptation in autonomous driving.
• Human language processing discerns the contextual use of words.

Numeric Predictions

• AI uses data for number-based forecasts or optimizations:
  • Weather Forecasting
  • Customer Subscription Renewal
  • Medical Risk Assessment
  • Sales Predictions
  • Pricing Optimization
  • Travel Expenses Optimization

Classifications

• AI identifies categories or labels for data:
  • Identifying Edible vs. Poisonous Plants
  • Email Legitimacy vs. Phishing
  • Financial Fraud Detection
  • Medical Diagnosis
  • Social Media Toxic Comment Identification

Robotic Navigation

• AI enables movement and decision-making for robots:
  • Autonomous Driving
  • Supply Chain Optimization
  • Adaptive Robots on Assembly Lines
  • Rescue Robots for Disaster Areas

Natural Language Processing (NLP)

• AI processes and generates human language:
  • Everyday Communication
  • ChatGPT for NLP
  • Translation Services
  • Summarizing Scientific Papers
  • Generative AI for Unique Images, Sounds, and Text