Skip to content

GenAI Use Cases

GenAI is transforming how we create, analyze, and discover - pushing the boundaries of what's possible across an astounding range of fields.

General Modalities

The following table provides an overview of the general modalities in which Generative AI can be applied:

Modality Examples
Language Spoken and Written
Time series Music, Speech, Finances
Visual 2D Images, Diagrams
Visual 3D 3D Models, Virtual Reality
Visual 2D with time Animated Graphics, Videos
Visual 3D with time 3D Animations, Simulations
Graphical Relation and Influence Networks
Generally linear sequences Genome, Proteome
Multidimensional Temporal sequences Weather, Brain Recordings, Stock Market
Multimodal variants Combination of the above methods

For a more detailed description of these modalities, refer to this section.

General Activities

Because at its core, GenAI works on Information, there are several fundamental ways in which Generative AI can be used. The application often depends on the field. Here are the core activities that can be used across many, if not all, fields of applications:

Creating Information

At its base, Generative AI is used to create information, such as new text or images. This creation can take several forms:

Expansion

  • Generating larger outputs from small inputs
  • Writing detailed documentation or articles
  • Brainstorming and ideation
  • Explaining complex concepts in detail
  • Creating training data for other AI systems

Reasoning

  • Evaluating trade-offs between different approaches
  • Analyzing complex scenarios and providing recommendations
  • Conducting risk assessments
  • Problem-solving with multiple variables
  • Strategic planning and decision-making

Converting Information

Generative AI can generate content in one domain with input from another. This includes:

  • Translating between languages (natural or programming)
  • Converting data formats (e.g., JSON to CSV)
  • Transforming natural language into structured queries
  • Converting visual information into textual descriptions
  • Transforming textual descriptions into visual representations

Compactifying Information

Generative AI excels at information compression and summarization:

  • Creating concise summaries of lengthy documents
  • Extracting key points from meetings or discussions
  • Distilling research papers into core findings
  • Generating executive summaries
  • Creating bullet-point highlights from detailed content
  • Even creating lossless compression at a fundamental level!
At a fundamental level Language Modeling Is Compression demonstrates 3x lossless compression of text and images.

Uses either newly trained 200K-3M transformer models or pre-trained Chinchilla models and achieves impressive compression rates. image Details on implementation are somewhat hidden.

Finding Information

Generative AI can understand and locate specific information:

  • Searching through documents for precise data points
  • Querying knowledge bases or databases
  • Finding relevant information in large datasets
  • Semantic search and relationship mapping
  • Answer extraction from complex documents

Taking Action

Generative AI can trigger and coordinate actions:

  • Generating executable commands
  • Orchestrating API calls
  • Managing workflow automation
  • Coordinating tool interactions
  • Implementing decision outcomes

Classifying and Predicting Information

While traditionally the domain of AI/ML, Generative AI can also perform:

  • Sentiment analysis and classification
  • Pattern recognition and prediction
  • Trend analysis and forecasting
  • Risk assessment and evaluation
  • Multi-label classification tasks

These activities can be combined to create more complex workflows, such as:

  • Finding relevant information, reasoning about it, and taking appropriate action
  • Converting information, compactifying it, and presenting insights
  • Creating new information based on patterns found in existing data